From e6a00f6684c9f348cd782922b4b277c68ab90b63 Mon Sep 17 00:00:00 2001 From: Yuval Shaia Date: Wed, 27 Jul 2016 01:24:52 -0700 Subject: IB/mlx4: Make function use_tunnel_data return void No need to return int if function always returns 0 Signed-off-by: Yuval Shaia Signed-off-by: Doug Ledford --- drivers/infiniband/hw/mlx4/cq.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/mlx4/cq.c b/drivers/infiniband/hw/mlx4/cq.c index d6fc8a6e8c33..15b628996633 100644 --- a/drivers/infiniband/hw/mlx4/cq.c +++ b/drivers/infiniband/hw/mlx4/cq.c @@ -576,8 +576,8 @@ static int mlx4_ib_ipoib_csum_ok(__be16 status, __be16 checksum) checksum == cpu_to_be16(0xffff); } -static int use_tunnel_data(struct mlx4_ib_qp *qp, struct mlx4_ib_cq *cq, struct ib_wc *wc, - unsigned tail, struct mlx4_cqe *cqe, int is_eth) +static void use_tunnel_data(struct mlx4_ib_qp *qp, struct mlx4_ib_cq *cq, struct ib_wc *wc, + unsigned tail, struct mlx4_cqe *cqe, int is_eth) { struct mlx4_ib_proxy_sqp_hdr *hdr; @@ -600,8 +600,6 @@ static int use_tunnel_data(struct mlx4_ib_qp *qp, struct mlx4_ib_cq *cq, struct wc->slid = be16_to_cpu(hdr->tun.slid_mac_47_32); wc->sl = (u8) (be16_to_cpu(hdr->tun.sl_vid) >> 12); } - - return 0; } static void mlx4_ib_qp_sw_comp(struct mlx4_ib_qp *qp, int num_entries, @@ -852,9 +850,11 @@ repoll: if (mlx4_is_mfunc(to_mdev(cq->ibcq.device)->dev)) { if ((*cur_qp)->mlx4_ib_qp_type & (MLX4_IB_QPT_PROXY_SMI_OWNER | - MLX4_IB_QPT_PROXY_SMI | MLX4_IB_QPT_PROXY_GSI)) - return use_tunnel_data(*cur_qp, cq, wc, tail, - cqe, is_eth); + MLX4_IB_QPT_PROXY_SMI | MLX4_IB_QPT_PROXY_GSI)) { + use_tunnel_data(*cur_qp, cq, wc, tail, cqe, + is_eth); + return 0; + } } wc->slid = be16_to_cpu(cqe->rlid); -- cgit From 5412352fcd8fba7f278ae8c9ba36296716f17ae8 Mon Sep 17 00:00:00 2001 From: Yuval Shaia Date: Wed, 27 Jul 2016 01:24:53 -0700 Subject: IB/mlx4: Return EAGAIN for any error in mlx4_ib_poll_one Error code EAGAIN should be used when errors are temporary and next call might succeeds. When error code other than EAGAIN is returned, the caller (mlx4_ib_poll) will assume all CQE in the same bunch are error too and will drop them all. Signed-off-by: Yuval Shaia Signed-off-by: Doug Ledford --- drivers/infiniband/hw/mlx4/cq.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/mlx4/cq.c b/drivers/infiniband/hw/mlx4/cq.c index 15b628996633..006db6436e3b 100644 --- a/drivers/infiniband/hw/mlx4/cq.c +++ b/drivers/infiniband/hw/mlx4/cq.c @@ -690,7 +690,7 @@ repoll: if (unlikely((cqe->owner_sr_opcode & MLX4_CQE_OPCODE_MASK) == MLX4_OPCODE_NOP && is_send)) { pr_warn("Completion for NOP opcode detected!\n"); - return -EINVAL; + return -EAGAIN; } /* Resize CQ in progress */ @@ -721,7 +721,7 @@ repoll: if (unlikely(!mqp)) { pr_warn("CQ %06x with entry for unknown QPN %06x\n", cq->mcq.cqn, be32_to_cpu(cqe->vlan_my_qpn) & MLX4_CQE_QPN_MASK); - return -EINVAL; + return -EAGAIN; } *cur_qp = to_mibqp(mqp); @@ -739,7 +739,7 @@ repoll: if (unlikely(!msrq)) { pr_warn("CQ %06x with entry for unknown SRQN %06x\n", cq->mcq.cqn, srq_num); - return -EINVAL; + return -EAGAIN; } } -- cgit From 8303f683b161467b6595c153c8751b80f9df3508 Mon Sep 17 00:00:00 2001 From: Tadeusz Struk Date: Wed, 3 Aug 2016 20:19:32 -0400 Subject: IB/hfi1: Allocate cpu mask on the heap to silence warning If CONFIG_FRAME_WARN is small (1K) and CONFIG_NR_CPUS big then a frame size warning is triggered during build. Allocate the cpu mask dynamically to silence the warning. Reviewed-by: Sebastian Sanchez Signed-off-by: Tadeusz Struk Signed-off-by: Doug Ledford --- drivers/infiniband/hw/hfi1/affinity.c | 20 +++++++++++++------- 1 file changed, 13 insertions(+), 7 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/hfi1/affinity.c b/drivers/infiniband/hw/hfi1/affinity.c index 79575ee873f2..9bbb21459166 100644 --- a/drivers/infiniband/hw/hfi1/affinity.c +++ b/drivers/infiniband/hw/hfi1/affinity.c @@ -682,7 +682,7 @@ int hfi1_set_sdma_affinity(struct hfi1_devdata *dd, const char *buf, size_t count) { struct hfi1_affinity_node *entry; - struct cpumask mask; + cpumask_var_t mask; int ret, i; spin_lock(&node_affinity.lock); @@ -692,19 +692,24 @@ int hfi1_set_sdma_affinity(struct hfi1_devdata *dd, const char *buf, if (!entry) return -EINVAL; - ret = cpulist_parse(buf, &mask); + ret = zalloc_cpumask_var(&mask, GFP_KERNEL); + if (!ret) + return -ENOMEM; + + ret = cpulist_parse(buf, mask); if (ret) - return ret; + goto out; - if (!cpumask_subset(&mask, cpu_online_mask) || cpumask_empty(&mask)) { + if (!cpumask_subset(mask, cpu_online_mask) || cpumask_empty(mask)) { dd_dev_warn(dd, "Invalid CPU mask\n"); - return -EINVAL; + ret = -EINVAL; + goto out; } mutex_lock(&sdma_affinity_mutex); /* reset the SDMA interrupt affinity details */ init_cpu_mask_set(&entry->def_intr); - cpumask_copy(&entry->def_intr.mask, &mask); + cpumask_copy(&entry->def_intr.mask, mask); /* * Reassign the affinity for each SDMA interrupt. */ @@ -720,8 +725,9 @@ int hfi1_set_sdma_affinity(struct hfi1_devdata *dd, const char *buf, if (ret) break; } - mutex_unlock(&sdma_affinity_mutex); +out: + free_cpumask_var(mask); return ret ? ret : strnlen(buf, PAGE_SIZE); } -- cgit From abb658ef0529488cfa75e79a2a9f894cb95eaec3 Mon Sep 17 00:00:00 2001 From: Wei Yongjun Date: Mon, 8 Aug 2016 09:49:47 +0000 Subject: IB/hfi1: Remove duplicated include from affinity.c Remove duplicated include. Signed-off-by: Wei Yongjun Reviewed-by: Leon Romanovsky Reviewed-by: Ira Weiny Signed-off-by: Doug Ledford --- drivers/infiniband/hw/hfi1/affinity.c | 1 - 1 file changed, 1 deletion(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/hfi1/affinity.c b/drivers/infiniband/hw/hfi1/affinity.c index 9bbb21459166..0566393e5aba 100644 --- a/drivers/infiniband/hw/hfi1/affinity.c +++ b/drivers/infiniband/hw/hfi1/affinity.c @@ -47,7 +47,6 @@ #include #include #include -#include #include "hfi.h" #include "affinity.h" -- cgit From c62fb260a86dde3df5b2905432caa0e9f6898434 Mon Sep 17 00:00:00 2001 From: Mike Marciniszyn Date: Fri, 12 Aug 2016 11:17:37 -0400 Subject: IB/hfi1,IB/qib: Fix qp_stats sleep with rcu read lock held The qp init function does a kzalloc() while holding the RCU lock that encounters the following warning with a debug kernel when a cat of the qp_stats is done: [ 231.723948] rcu_scheduler_active = 1, debug_locks = 0 [ 231.731939] 3 locks held by cat/11355: [ 231.736492] #0: (debugfs_srcu){......}, at: [] debugfs_use_file_start+0x5/0x90 [ 231.746955] #1: (&p->lock){+.+.+.}, at: [] seq_read+0x4c/0x3c0 [ 231.755873] #2: (rcu_read_lock){......}, at: [] _qp_stats_seq_start+0x5/0xd0 [hfi1] [ 231.766862] The init functions do an implicit next which requires the rcu read lock before the kzalloc(). Fix for both drivers is to change the scope of the init function to only do the allocation and the initialization of the just allocated iter. The implict next is moved back into the respective start functions to fix the issue. Signed-off-by: Ira Weiny Signed-off-by: Mike Marciniszyn CC: # 4.6.x- Reviewed-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/hw/hfi1/debugfs.c | 14 +++++++++----- drivers/infiniband/hw/hfi1/qp.c | 4 ---- drivers/infiniband/hw/qib/qib_debugfs.c | 12 +++++++++--- drivers/infiniband/hw/qib/qib_qp.c | 4 ---- 4 files changed, 18 insertions(+), 16 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/hfi1/debugfs.c b/drivers/infiniband/hw/hfi1/debugfs.c index dbab9d9cc288..a49cc88f08a2 100644 --- a/drivers/infiniband/hw/hfi1/debugfs.c +++ b/drivers/infiniband/hw/hfi1/debugfs.c @@ -223,28 +223,32 @@ DEBUGFS_SEQ_FILE_OPEN(ctx_stats) DEBUGFS_FILE_OPS(ctx_stats); static void *_qp_stats_seq_start(struct seq_file *s, loff_t *pos) -__acquires(RCU) + __acquires(RCU) { struct qp_iter *iter; loff_t n = *pos; - rcu_read_lock(); iter = qp_iter_init(s->private); + + /* stop calls rcu_read_unlock */ + rcu_read_lock(); + if (!iter) return NULL; - while (n--) { + do { if (qp_iter_next(iter)) { kfree(iter); return NULL; } - } + } while (n--); return iter; } static void *_qp_stats_seq_next(struct seq_file *s, void *iter_ptr, loff_t *pos) + __must_hold(RCU) { struct qp_iter *iter = iter_ptr; @@ -259,7 +263,7 @@ static void *_qp_stats_seq_next(struct seq_file *s, void *iter_ptr, } static void _qp_stats_seq_stop(struct seq_file *s, void *iter_ptr) -__releases(RCU) + __releases(RCU) { rcu_read_unlock(); } diff --git a/drivers/infiniband/hw/hfi1/qp.c b/drivers/infiniband/hw/hfi1/qp.c index a5aa3517e7d5..4e4d8317c281 100644 --- a/drivers/infiniband/hw/hfi1/qp.c +++ b/drivers/infiniband/hw/hfi1/qp.c @@ -656,10 +656,6 @@ struct qp_iter *qp_iter_init(struct hfi1_ibdev *dev) iter->dev = dev; iter->specials = dev->rdi.ibdev.phys_port_cnt * 2; - if (qp_iter_next(iter)) { - kfree(iter); - return NULL; - } return iter; } diff --git a/drivers/infiniband/hw/qib/qib_debugfs.c b/drivers/infiniband/hw/qib/qib_debugfs.c index 5e75b43c596b..5bad8e3b40bb 100644 --- a/drivers/infiniband/hw/qib/qib_debugfs.c +++ b/drivers/infiniband/hw/qib/qib_debugfs.c @@ -189,27 +189,32 @@ static int _ctx_stats_seq_show(struct seq_file *s, void *v) DEBUGFS_FILE(ctx_stats) static void *_qp_stats_seq_start(struct seq_file *s, loff_t *pos) + __acquires(RCU) { struct qib_qp_iter *iter; loff_t n = *pos; - rcu_read_lock(); iter = qib_qp_iter_init(s->private); + + /* stop calls rcu_read_unlock */ + rcu_read_lock(); + if (!iter) return NULL; - while (n--) { + do { if (qib_qp_iter_next(iter)) { kfree(iter); return NULL; } - } + } while (n--); return iter; } static void *_qp_stats_seq_next(struct seq_file *s, void *iter_ptr, loff_t *pos) + __must_hold(RCU) { struct qib_qp_iter *iter = iter_ptr; @@ -224,6 +229,7 @@ static void *_qp_stats_seq_next(struct seq_file *s, void *iter_ptr, } static void _qp_stats_seq_stop(struct seq_file *s, void *iter_ptr) + __releases(RCU) { rcu_read_unlock(); } diff --git a/drivers/infiniband/hw/qib/qib_qp.c b/drivers/infiniband/hw/qib/qib_qp.c index 9cc0aae1d781..f9b8cd2354d1 100644 --- a/drivers/infiniband/hw/qib/qib_qp.c +++ b/drivers/infiniband/hw/qib/qib_qp.c @@ -573,10 +573,6 @@ struct qib_qp_iter *qib_qp_iter_init(struct qib_ibdev *dev) return NULL; iter->dev = dev; - if (qib_qp_iter_next(iter)) { - kfree(iter); - return NULL; - } return iter; } -- cgit From 140690eae7626a820ea135f33ff9fe11c1bcee6d Mon Sep 17 00:00:00 2001 From: Easwar Hariharan Date: Tue, 9 Aug 2016 11:17:18 -0400 Subject: IB/hfi1: Fetch monitor values on-demand for CableInfo query The monitor values from bytes 22 through 81 of the QSFP memory space (SFF 8636) are dynamic and serving them out of the QSFP memory cache maintained by the driver provides stale data to the CableInfo SMA query. This patch refreshes the dynamic values from the QSFP memory on request and overwrites the stale data from the cache for the overlap between the requested range and the monitor range. Reviewed-by: Jubin John Reviewed-by: Ira Weiny Signed-off-by: Easwar Hariharan Signed-off-by: Doug Ledford --- drivers/infiniband/hw/hfi1/qsfp.c | 32 ++++++++++++++++++++++++++++++-- drivers/infiniband/hw/hfi1/qsfp.h | 3 +++ 2 files changed, 33 insertions(+), 2 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/hfi1/qsfp.c b/drivers/infiniband/hw/hfi1/qsfp.c index a207717ade2a..4e95ad810847 100644 --- a/drivers/infiniband/hw/hfi1/qsfp.c +++ b/drivers/infiniband/hw/hfi1/qsfp.c @@ -706,8 +706,8 @@ int get_cable_info(struct hfi1_devdata *dd, u32 port_num, u32 addr, u32 len, u8 *data) { struct hfi1_pportdata *ppd; - u32 excess_len = 0; - int ret = 0; + u32 excess_len = len; + int ret = 0, offset = 0; if (port_num > dd->num_pports || port_num < 1) { dd_dev_info(dd, "%s: Invalid port number %d\n", @@ -740,6 +740,34 @@ int get_cable_info(struct hfi1_devdata *dd, u32 port_num, u32 addr, u32 len, } memcpy(data, &ppd->qsfp_info.cache[addr], len); + + if (addr <= QSFP_MONITOR_VAL_END && + (addr + len) >= QSFP_MONITOR_VAL_START) { + /* Overlap with the dynamic channel monitor range */ + if (addr < QSFP_MONITOR_VAL_START) { + if (addr + len <= QSFP_MONITOR_VAL_END) + len = addr + len - QSFP_MONITOR_VAL_START; + else + len = QSFP_MONITOR_RANGE; + offset = QSFP_MONITOR_VAL_START - addr; + addr = QSFP_MONITOR_VAL_START; + } else if (addr == QSFP_MONITOR_VAL_START) { + offset = 0; + if (addr + len > QSFP_MONITOR_VAL_END) + len = QSFP_MONITOR_RANGE; + } else { + offset = 0; + if (addr + len > QSFP_MONITOR_VAL_END) + len = QSFP_MONITOR_VAL_END - addr + 1; + } + /* Refresh the values of the dynamic monitors from the cable */ + ret = one_qsfp_read(ppd, dd->hfi1_id, addr, data + offset, len); + if (ret != len) { + ret = -EAGAIN; + goto set_zeroes; + } + } + return 0; set_zeroes: diff --git a/drivers/infiniband/hw/hfi1/qsfp.h b/drivers/infiniband/hw/hfi1/qsfp.h index 69275ebd9597..36cf52359848 100644 --- a/drivers/infiniband/hw/hfi1/qsfp.h +++ b/drivers/infiniband/hw/hfi1/qsfp.h @@ -74,6 +74,9 @@ /* Defined fields that Intel requires of qualified cables */ /* Byte 0 is Identifier, not checked */ /* Byte 1 is reserved "status MSB" */ +#define QSFP_MONITOR_VAL_START 22 +#define QSFP_MONITOR_VAL_END 81 +#define QSFP_MONITOR_RANGE (QSFP_MONITOR_VAL_END - QSFP_MONITOR_VAL_START + 1) #define QSFP_TX_CTRL_BYTE_OFFS 86 #define QSFP_PWR_CTRL_BYTE_OFFS 93 #define QSFP_CDR_CTRL_BYTE_OFFS 98 -- cgit From c867caaf8ecd5498d399558d0f11825f175c6cdd Mon Sep 17 00:00:00 2001 From: Mike Marciniszyn Date: Tue, 9 Aug 2016 11:19:55 -0400 Subject: IB/hfi1: Pass packet ptr to set_armed_active The "packet" parameter was being passed on the stack, change it to a pointer. Reviewed-by: Don Hiatt Signed-off-by: Mike Marciniszyn Signed-off-by: Doug Ledford --- drivers/infiniband/hw/hfi1/driver.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/hfi1/driver.c b/drivers/infiniband/hw/hfi1/driver.c index 8246dc7d0573..c4156870a8bb 100644 --- a/drivers/infiniband/hw/hfi1/driver.c +++ b/drivers/infiniband/hw/hfi1/driver.c @@ -888,14 +888,14 @@ void set_all_slowpath(struct hfi1_devdata *dd) } static inline int set_armed_to_active(struct hfi1_ctxtdata *rcd, - struct hfi1_packet packet, + struct hfi1_packet *packet, struct hfi1_devdata *dd) { struct work_struct *lsaw = &rcd->ppd->linkstate_active_work; - struct hfi1_message_header *hdr = hfi1_get_msgheader(packet.rcd->dd, - packet.rhf_addr); + struct hfi1_message_header *hdr = hfi1_get_msgheader(packet->rcd->dd, + packet->rhf_addr); - if (hdr2sc(hdr, packet.rhf) != 0xf) { + if (hdr2sc(hdr, packet->rhf) != 0xf) { int hwstate = read_logical_state(dd); if (hwstate != LSTATE_ACTIVE) { @@ -979,7 +979,7 @@ int handle_receive_interrupt(struct hfi1_ctxtdata *rcd, int thread) /* Auto activate link on non-SC15 packet receive */ if (unlikely(rcd->ppd->host_link_state == HLS_UP_ARMED) && - set_armed_to_active(rcd, packet, dd)) + set_armed_to_active(rcd, &packet, dd)) goto bail; last = process_rcv_packet(&packet, thread); } -- cgit From 69b9f4a4233b6f9ea168c25123b9eb629739d8e5 Mon Sep 17 00:00:00 2001 From: Mike Marciniszyn Date: Tue, 9 Aug 2016 11:19:56 -0400 Subject: IB/hfi1: Validate header in set_armed_active Validate the etype to insure that the header is correct. Reviewed-by: Don Hiatt Signed-off-by: Mike Marciniszyn Signed-off-by: Doug Ledford --- drivers/infiniband/hw/hfi1/driver.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/hfi1/driver.c b/drivers/infiniband/hw/hfi1/driver.c index c4156870a8bb..303f10555729 100644 --- a/drivers/infiniband/hw/hfi1/driver.c +++ b/drivers/infiniband/hw/hfi1/driver.c @@ -894,8 +894,9 @@ static inline int set_armed_to_active(struct hfi1_ctxtdata *rcd, struct work_struct *lsaw = &rcd->ppd->linkstate_active_work; struct hfi1_message_header *hdr = hfi1_get_msgheader(packet->rcd->dd, packet->rhf_addr); + u8 etype = rhf_rcv_type(packet->rhf); - if (hdr2sc(hdr, packet->rhf) != 0xf) { + if (etype == RHF_RCV_TYPE_IB && hdr2sc(hdr, packet->rhf) != 0xf) { int hwstate = read_logical_state(dd); if (hwstate != LSTATE_ACTIVE) { -- cgit From 476d95bd02240894806ebe64f1c4dcf6dbba87f4 Mon Sep 17 00:00:00 2001 From: Wei Yongjun Date: Wed, 10 Aug 2016 03:14:04 +0000 Subject: IB/hfi1: Using kfree_rcu() to simplify the code The callback function of call_rcu() just calls a kfree(), so we can use kfree_rcu() instead of call_rcu() + callback function. Signed-off-by: Wei Yongjun Tested-by: Mike Marciniszyn Acked-by: Mike Marciniszyn Tested-by: Mike Marciniszyn Acked-by: Mike Marciniszyn Signed-off-by: Doug Ledford --- drivers/infiniband/hw/hfi1/hfi.h | 1 - drivers/infiniband/hw/hfi1/init.c | 2 +- drivers/infiniband/hw/hfi1/mad.c | 9 +-------- 3 files changed, 2 insertions(+), 10 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/hfi1/hfi.h b/drivers/infiniband/hw/hfi1/hfi.h index 1000e0fd96d9..f41414ef0cbc 100644 --- a/drivers/infiniband/hw/hfi1/hfi.h +++ b/drivers/infiniband/hw/hfi1/hfi.h @@ -1656,7 +1656,6 @@ struct cc_state *get_cc_state_protected(struct hfi1_pportdata *ppd) struct hfi1_devdata *hfi1_init_dd(struct pci_dev *, const struct pci_device_id *); void hfi1_free_devdata(struct hfi1_devdata *); -void cc_state_reclaim(struct rcu_head *rcu); struct hfi1_devdata *hfi1_alloc_devdata(struct pci_dev *pdev, size_t extra); /* LED beaconing functions */ diff --git a/drivers/infiniband/hw/hfi1/init.c b/drivers/infiniband/hw/hfi1/init.c index a358d23ecd54..b7935451093c 100644 --- a/drivers/infiniband/hw/hfi1/init.c +++ b/drivers/infiniband/hw/hfi1/init.c @@ -1333,7 +1333,7 @@ static void cleanup_device_data(struct hfi1_devdata *dd) spin_unlock(&ppd->cc_state_lock); if (cc_state) - call_rcu(&cc_state->rcu, cc_state_reclaim); + kfree_rcu(cc_state, rcu); } free_credit_return(dd); diff --git a/drivers/infiniband/hw/hfi1/mad.c b/drivers/infiniband/hw/hfi1/mad.c index 1263abe01999..95c43e1e0ce5 100644 --- a/drivers/infiniband/hw/hfi1/mad.c +++ b/drivers/infiniband/hw/hfi1/mad.c @@ -3398,7 +3398,7 @@ static void apply_cc_state(struct hfi1_pportdata *ppd) spin_unlock(&ppd->cc_state_lock); - call_rcu(&old_cc_state->rcu, cc_state_reclaim); + kfree_rcu(old_cc_state, rcu); } static int __subn_set_opa_cong_setting(struct opa_smp *smp, u32 am, u8 *data, @@ -3553,13 +3553,6 @@ static int __subn_get_opa_cc_table(struct opa_smp *smp, u32 am, u8 *data, return reply((struct ib_mad_hdr *)smp); } -void cc_state_reclaim(struct rcu_head *rcu) -{ - struct cc_state *cc_state = container_of(rcu, struct cc_state, rcu); - - kfree(cc_state); -} - static int __subn_set_opa_cc_table(struct opa_smp *smp, u32 am, u8 *data, struct ib_device *ibdev, u8 port, u32 *resp_len) -- cgit From 57bb562ad4651453c9f20d506a65e46e4d11042f Mon Sep 17 00:00:00 2001 From: Christophe Jaillet Date: Wed, 10 Aug 2016 07:34:27 +0200 Subject: IB/hfi1: Add missing error code assignment before test It is likely that checking the result of 'setup_ctxt' is expected here. Signed-off-by: Christophe JAILLET Acked-by: Dennis Dalessandro Signed-off-by: Doug Ledford --- drivers/infiniband/hw/hfi1/file_ops.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/hfi1/file_ops.c b/drivers/infiniband/hw/hfi1/file_ops.c index 1ecbec192358..ed76be39b20b 100644 --- a/drivers/infiniband/hw/hfi1/file_ops.c +++ b/drivers/infiniband/hw/hfi1/file_ops.c @@ -222,7 +222,7 @@ static long hfi1_file_ioctl(struct file *fp, unsigned int cmd, ret = assign_ctxt(fp, &uinfo); if (ret < 0) return ret; - setup_ctxt(fp); + ret = setup_ctxt(fp); if (ret) return ret; ret = user_init(fp); -- cgit From 86cd747c6dd6cdbc825e36ad5f0029f3c5a37776 Mon Sep 17 00:00:00 2001 From: Christophe Jaillet Date: Wed, 10 Aug 2016 17:45:01 +0200 Subject: IB/usnic: Fix error return code If 'pci_register_driver' fails, we return 'err' which is known to be 0. Return the error instead. Signed-off-by: Christophe JAILLET Signed-off-by: Doug Ledford --- drivers/infiniband/hw/usnic/usnic_ib_main.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/usnic/usnic_ib_main.c b/drivers/infiniband/hw/usnic/usnic_ib_main.c index c229b9f4a52d..0a89a955550b 100644 --- a/drivers/infiniband/hw/usnic/usnic_ib_main.c +++ b/drivers/infiniband/hw/usnic/usnic_ib_main.c @@ -664,7 +664,8 @@ static int __init usnic_ib_init(void) return err; } - if (pci_register_driver(&usnic_ib_pci_driver)) { + err = pci_register_driver(&usnic_ib_pci_driver); + if (err) { usnic_err("Unable to register with PCI\n"); goto out_umem_fini; } -- cgit From f29a08dc145e05de6a57f0aeaba6020464f80e15 Mon Sep 17 00:00:00 2001 From: Easwar Hariharan Date: Tue, 16 Aug 2016 13:25:34 -0700 Subject: IB/hfi1: Return invalid field for non-QSFP CableInfo queries The driver does not check if the CableInfo query is supported for the port type. Return early if CableInfo is not supported for the port type, making compliance with the specification explicit and preventing lower level code from potentially doing the wrong thing if the query is not supported for the hardware implementation. Reviewed-by: Ira Weiny Signed-off-by: Easwar Hariharan Signed-off-by: Dennis Dalessandro Signed-off-by: Doug Ledford --- drivers/infiniband/hw/hfi1/mad.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/hfi1/mad.c b/drivers/infiniband/hw/hfi1/mad.c index 95c43e1e0ce5..39e42c373a01 100644 --- a/drivers/infiniband/hw/hfi1/mad.c +++ b/drivers/infiniband/hw/hfi1/mad.c @@ -1819,6 +1819,11 @@ static int __subn_get_opa_cable_info(struct opa_smp *smp, u32 am, u8 *data, u32 len = OPA_AM_CI_LEN(am) + 1; int ret; + if (dd->pport->port_type != PORT_TYPE_QSFP) { + smp->status |= IB_SMP_INVALID_FIELD; + return reply((struct ib_mad_hdr *)smp); + } + #define __CI_PAGE_SIZE BIT(7) /* 128 bytes */ #define __CI_PAGE_MASK ~(__CI_PAGE_SIZE - 1) #define __CI_PAGE_NUM(a) ((a) & __CI_PAGE_MASK) -- cgit From 08fe16f6192bccd5798e9b60461f7aa151b34cd4 Mon Sep 17 00:00:00 2001 From: Mitko Haralanov Date: Tue, 16 Aug 2016 13:26:12 -0700 Subject: IB/hfi1: Improve J_KEY generation Previously, J_KEY generation was based on the lower 16 bits of the user's UID. While this works, it was not good enough as a non-root user could collide with a root user given a sufficiently large UID. This patch attempt to improve the J_KEY generation by using the following algorithm: The 16 bit J_KEY space is partitioned into 3 separate spaces reserved for different user classes: * all users with administtor privileges (including 'root') will use J_KEYs in the range of 0 to 31, * all kernel protocols, which use KDETH packets will use J_KEYs in the range of 32 to 63, and * all other users will use J_KEYs in the range of 64 to 65535. The above separation is aimed at preventing different user levels from sending packets to each other and, additionally, separate kernel protocols from all other types of users. The later is meant to prevent the potential corruption of kernel memory by any other type of user. Reviewed-by: Ira Weiny Signed-off-by: Mitko Haralanov Signed-off-by: Dennis Dalessandro Signed-off-by: Doug Ledford --- drivers/infiniband/hw/hfi1/hfi.h | 19 ++++++++++++++++++- 1 file changed, 18 insertions(+), 1 deletion(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/hfi1/hfi.h b/drivers/infiniband/hw/hfi1/hfi.h index f41414ef0cbc..a021e660d482 100644 --- a/drivers/infiniband/hw/hfi1/hfi.h +++ b/drivers/infiniband/hw/hfi1/hfi.h @@ -1272,9 +1272,26 @@ static inline int hdr2sc(struct hfi1_message_header *hdr, u64 rhf) ((!!(rhf_dc_info(rhf))) << 4); } +#define HFI1_JKEY_WIDTH 16 +#define HFI1_JKEY_MASK (BIT(16) - 1) +#define HFI1_ADMIN_JKEY_RANGE 32 + +/* + * J_KEYs are split and allocated in the following groups: + * 0 - 31 - users with administrator privileges + * 32 - 63 - kernel protocols using KDETH packets + * 64 - 65535 - all other users using KDETH packets + */ static inline u16 generate_jkey(kuid_t uid) { - return from_kuid(current_user_ns(), uid) & 0xffff; + u16 jkey = from_kuid(current_user_ns(), uid) & HFI1_JKEY_MASK; + + if (capable(CAP_SYS_ADMIN)) + jkey &= HFI1_ADMIN_JKEY_RANGE - 1; + else if (jkey < 64) + jkey |= BIT(HFI1_JKEY_WIDTH - 1); + + return jkey; } /* -- cgit From e0cf75deab8155334c8228eb7f097b15127d0a49 Mon Sep 17 00:00:00 2001 From: Ira Weiny Date: Tue, 16 Aug 2016 13:27:03 -0700 Subject: IB/hfi1: Fix mm_struct use after free Testing with CONFIG_SLUB_DEBUG_ON=y resulted in the kernel panic below. This is the result of the mm_struct sometimes being free'd prior to hfi1_file_close being called. This was due to the combination of 2 reasons: 1) hfi1_file_close is deferred in process exit and it therefore may not be called synchronously with process exit. 2) exit_mm is called prior to exit_files in do_exit. Normally this is ok however, our kernel bypass code requires us to have access to the mm_struct for house keeping both at "normal" close time as well as at process exit. Therefore, the fix is to simply keep a reference to the mm_struct until we are done with it. [ 3006.340150] general protection fault: 0000 [#1] SMP [ 3006.346469] Modules linked in: hfi1 rdmavt rpcrdma ib_isert iscsi_target_mod ib_iser libiscsi scsi_transport_iscsi ib_srpt target_core_mod ib_srp scsi_transport_srp ib_ipoib rdma_ucm ib_ucm ib_uverbs ib_umad rdma_cm ib_cm iw_cm dm_mirror dm_region_hash dm_log dm_mod snd_hda_code c_realtek iTCO_wdt snd_hda_codec_generic iTCO_vendor_support sb_edac edac_core x86_pkg_temp_thermal intel_powerclamp coretemp kvm irqbypass c rct10dif_pclmul crc32_pclmul ghash_clmulni_intel aesni_intel lrw snd_hda_intel gf128mul snd_hda_codec glue_helper snd_hda_core ablk_helper sn d_hwdep cryptd snd_seq snd_seq_device snd_pcm snd_timer snd soundcore pcspkr shpchp mei_me sg lpc_ich mei i2c_i801 mfd_core ioatdma ipmi_devi ntf wmi ipmi_si ipmi_msghandler acpi_cpufreq nfsd auth_rpcgss nfs_acl lockd grace sunrpc ip_tables ext4 jbd2 mbcache mlx4_en ib_core sr_mod s d_mod cdrom crc32c_intel mgag200 drm_kms_helper syscopyarea sysfillrect igb sysimgblt fb_sys_fops ptp mlx4_core ttm isci pps_core ahci drm li bsas libahci dca firewire_ohci i2c_algo_bit scsi_transport_sas firewire_core crc_itu_t i2c_core libata [last unloaded: mlx4_ib] [ 3006.461759] CPU: 16 PID: 11624 Comm: mpi_stress Not tainted 4.7.0-rc5+ #1 [ 3006.469915] Hardware name: Intel Corporation W2600CR ........../W2600CR, BIOS SE5C600.86B.01.08.0003.022620131521 02/26/2013 [ 3006.483027] task: ffff8804102f0040 ti: ffff8804102f8000 task.ti: ffff8804102f8000 [ 3006.491971] RIP: 0010:[] [] __lock_acquire+0xb3/0x19e0 [ 3006.501905] RSP: 0018:ffff8804102fb908 EFLAGS: 00010002 [ 3006.508447] RAX: 6b6b6b6b6b6b6b6b RBX: 0000000000000001 RCX: 0000000000000000 [ 3006.517012] RDX: 0000000000000001 RSI: 0000000000000000 RDI: ffff880410b56a40 [ 3006.525569] RBP: ffff8804102fb9b0 R08: 0000000000000001 R09: 0000000000000000 [ 3006.534119] R10: ffff8804102f0040 R11: 0000000000000000 R12: 0000000000000000 [ 3006.542664] R13: ffff880410b56a40 R14: 0000000000000000 R15: 0000000000000000 [ 3006.551203] FS: 00007ff478c08700(0000) GS:ffff88042e200000(0000) knlGS:0000000000000000 [ 3006.560814] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 3006.567806] CR2: 00007f667f5109e0 CR3: 0000000001c06000 CR4: 00000000000406e0 [ 3006.576352] Stack: [ 3006.579157] ffffffff8124b819 ffffffffffffffff 0000000000000000 ffff8804102fb940 [ 3006.588072] 0000000000000002 0000000000000000 ffff8804102f0040 0000000000000007 [ 3006.596971] 0000000000000006 ffff8803cad6f000 0000000000000000 ffff8804102f0040 [ 3006.605878] Call Trace: [ 3006.609220] [] ? uncharge_batch+0x109/0x250 [ 3006.616382] [] lock_acquire+0xd3/0x220 [ 3006.623056] [] ? hfi1_release_user_pages+0x7c/0xa0 [hfi1] [ 3006.631593] [] down_write+0x49/0x80 [ 3006.638022] [] ? hfi1_release_user_pages+0x7c/0xa0 [hfi1] [ 3006.646569] [] hfi1_release_user_pages+0x7c/0xa0 [hfi1] [ 3006.654898] [] cacheless_tid_rb_remove+0x106/0x330 [hfi1] [ 3006.663417] [] ? mark_held_locks+0x66/0x90 [ 3006.670498] [] ? _raw_spin_unlock_irqrestore+0x36/0x60 [ 3006.678741] [] tid_rb_remove+0xe/0x10 [hfi1] [ 3006.686010] [] hfi1_mmu_rb_unregister+0xc5/0x100 [hfi1] [ 3006.694387] [] hfi1_user_exp_rcv_free+0x39/0x120 [hfi1] [ 3006.702732] [] hfi1_file_close+0x17a/0x330 [hfi1] [ 3006.710489] [] __fput+0xfa/0x230 [ 3006.716595] [] ____fput+0xe/0x10 [ 3006.722696] [] task_work_run+0x86/0xc0 [ 3006.729379] [] do_exit+0x323/0xc40 [ 3006.735672] [] do_group_exit+0x4c/0xc0 [ 3006.742371] [] get_signal+0x345/0x940 [ 3006.748958] [] do_signal+0x37/0x700 [ 3006.755328] [] ? poll_select_set_timeout+0x5a/0x90 [ 3006.763146] [] ? __audit_syscall_exit+0x1db/0x260 [ 3006.770853] [] ? rcu_read_lock_sched_held+0x93/0xa0 [ 3006.778765] [] ? kfree+0x1e4/0x2a0 [ 3006.784986] [] ? exit_to_usermode_loop+0x33/0xac [ 3006.792551] [] exit_to_usermode_loop+0x5e/0xac [ 3006.799907] [] do_syscall_64+0x12a/0x190 [ 3006.806664] [] entry_SYSCALL64_slow_path+0x25/0x25 [ 3006.814396] Code: 24 08 44 89 44 24 10 89 4c 24 18 e8 a8 d8 ff ff 48 85 c0 8b 4c 24 18 44 8b 44 24 10 44 8b 4c 24 08 4c 8b 14 24 0f 84 30 08 00 00 ff 80 98 01 00 00 8b 3d 48 ad be 01 45 8b a2 90 0b 00 00 85 [ 3006.837158] RIP [] __lock_acquire+0xb3/0x19e0 [ 3006.844401] RSP [ 3006.851170] ---[ end trace b7b9f21cf06c27df ]--- [ 3006.927420] Kernel panic - not syncing: Fatal exception [ 3006.933954] Kernel Offset: disabled [ 3006.940961] ---[ end Kernel panic - not syncing: Fatal exception [ 3006.948249] ------------[ cut here ]------------ Fixes: 3faa3d9a308e ("IB/hfi1: Make use of mm consistent") Reviewed-by: Dean Luick Signed-off-by: Ira Weiny Signed-off-by: Dennis Dalessandro Signed-off-by: Doug Ledford --- drivers/infiniband/hw/hfi1/file_ops.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/hfi1/file_ops.c b/drivers/infiniband/hw/hfi1/file_ops.c index ed76be39b20b..7e03ccd2554d 100644 --- a/drivers/infiniband/hw/hfi1/file_ops.c +++ b/drivers/infiniband/hw/hfi1/file_ops.c @@ -183,6 +183,7 @@ static int hfi1_file_open(struct inode *inode, struct file *fp) if (fd) { fd->rec_cpu_num = -1; /* no cpu affinity by default */ fd->mm = current->mm; + atomic_inc(&fd->mm->mm_count); } fp->private_data = fd; @@ -779,6 +780,7 @@ static int hfi1_file_close(struct inode *inode, struct file *fp) mutex_unlock(&hfi1_mutex); hfi1_free_ctxtdata(dd, uctxt); done: + mmdrop(fdata->mm); kobject_put(&dd->kobj); kfree(fdata); return 0; -- cgit From 7f446abf12d741f4e29f00c633cb5fa67c2eee71 Mon Sep 17 00:00:00 2001 From: Steve Wise Date: Fri, 19 Aug 2016 07:29:07 -0700 Subject: iw_cxgb4: limit IRD/ORD advertised to ULP by device max. The i40iw initiator sends an MPA-request with ird = 63, ord = 63. The cxgb4 responder sends a RST. Since the inbound ord=63 and it exceeds the max_ird/c4iw_max_read_depth (=32 default), chelsio decides to abort. Instead, cxgb4 should adjust the ord/ird down before presenting it to the ULP. Reported-by: Shiraz Saleem Signed-off-by: Steve Wise Signed-off-by: Doug Ledford --- drivers/infiniband/hw/cxgb4/cm.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/cxgb4/cm.c b/drivers/infiniband/hw/cxgb4/cm.c index 3aca7f6171b4..a698efc69257 100644 --- a/drivers/infiniband/hw/cxgb4/cm.c +++ b/drivers/infiniband/hw/cxgb4/cm.c @@ -1827,8 +1827,12 @@ static int process_mpa_request(struct c4iw_ep *ep, struct sk_buff *skb) (ep->mpa_pkt + sizeof(*mpa)); ep->ird = ntohs(mpa_v2_params->ird) & MPA_V2_IRD_ORD_MASK; + ep->ird = min_t(u32, ep->ird, + cur_max_read_depth(ep->com.dev)); ep->ord = ntohs(mpa_v2_params->ord) & MPA_V2_IRD_ORD_MASK; + ep->ord = min_t(u32, ep->ord, + cur_max_read_depth(ep->com.dev)); PDBG("%s initiator ird %u ord %u\n", __func__, ep->ird, ep->ord); if (ntohs(mpa_v2_params->ird) & MPA_V2_PEER2PEER_MODEL) -- cgit From 30b03b1528b60623986ee0b50ec0e5dab9094be8 Mon Sep 17 00:00:00 2001 From: Steve Wise Date: Fri, 19 Aug 2016 07:29:08 -0700 Subject: iw_cxgb4: use the MPA initiator's IRD if < our ORD The i40iw initiator sends an MPA-request with ird=16 and ord=16. The cxgb4 responder sends an MPA-reply with ord = 32 causing i40iw to terminate due to insufficient resources. The logic to reduce the ORD to <= peer's IRD was wrong. Reported-by: Shiraz Saleem Signed-off-by: Steve Wise Signed-off-by: Doug Ledford --- drivers/infiniband/hw/cxgb4/cm.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/cxgb4/cm.c b/drivers/infiniband/hw/cxgb4/cm.c index a698efc69257..b6a953aed7e8 100644 --- a/drivers/infiniband/hw/cxgb4/cm.c +++ b/drivers/infiniband/hw/cxgb4/cm.c @@ -3140,7 +3140,7 @@ int c4iw_accept_cr(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) if (ep->mpa_attr.version == 2 && ep->mpa_attr.enhanced_rdma_conn) { if (conn_param->ord > ep->ird) { if (RELAXED_IRD_NEGOTIATION) { - ep->ord = ep->ird; + conn_param->ord = ep->ird; } else { ep->ird = conn_param->ird; ep->ord = conn_param->ord; -- cgit From 48ef5865d08fa0a36d786f2f8e12c6194d27538b Mon Sep 17 00:00:00 2001 From: Markus Elfring Date: Fri, 19 Aug 2016 08:50:23 +0200 Subject: IB/qib: Use memdup_user() rather than duplicating its implementation Reuse existing functionality from memdup_user() instead of keeping duplicate source code. This issue was detected by using the Coccinelle software. Signed-off-by: Markus Elfring Reviewed-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/hw/qib/qib_fs.c | 26 +++++--------------------- 1 file changed, 5 insertions(+), 21 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/qib/qib_fs.c b/drivers/infiniband/hw/qib/qib_fs.c index fcdf37913a26..c3edc033f7c4 100644 --- a/drivers/infiniband/hw/qib/qib_fs.c +++ b/drivers/infiniband/hw/qib/qib_fs.c @@ -328,26 +328,12 @@ static ssize_t flash_write(struct file *file, const char __user *buf, pos = *ppos; - if (pos != 0) { - ret = -EINVAL; - goto bail; - } - - if (count != sizeof(struct qib_flash)) { - ret = -EINVAL; - goto bail; - } - - tmp = kmalloc(count, GFP_KERNEL); - if (!tmp) { - ret = -ENOMEM; - goto bail; - } + if (pos != 0 || count != sizeof(struct qib_flash)) + return -EINVAL; - if (copy_from_user(tmp, buf, count)) { - ret = -EFAULT; - goto bail_tmp; - } + tmp = memdup_user(buf, count); + if (IS_ERR(tmp)) + return PTR_ERR(tmp); dd = private2dd(file); if (qib_eeprom_write(dd, pos, tmp, count)) { @@ -361,8 +347,6 @@ static ssize_t flash_write(struct file *file, const char __user *buf, bail_tmp: kfree(tmp); - -bail: return ret; } -- cgit From 6c7d46fdb8165ece4b0a17fb8f0b9320dbfeffc2 Mon Sep 17 00:00:00 2001 From: Shiraz Saleem Date: Mon, 22 Aug 2016 18:09:14 -0500 Subject: i40iw: Change mem_resources pointer to a u8 iwdev->mem_resources is incorrectly defined as an unsigned long instead of u8. As a result, the offset into the dynamic allocated structures in i40iw_initialize_hw_resources() is incorrectly calculated and would lead to writing of memory regions outside of the allocated buffer. Fixes: 8e06af711bf2 ("i40iw: add main, hdr, status") Reported-by: Stefan Assmann Signed-off-by: Mustafa Ismail Signed-off-by: Shiraz Saleem Signed-off-by: Doug Ledford --- drivers/infiniband/hw/i40iw/i40iw.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/i40iw/i40iw.h b/drivers/infiniband/hw/i40iw/i40iw.h index b738acdb9b02..882f3ef2e4c7 100644 --- a/drivers/infiniband/hw/i40iw/i40iw.h +++ b/drivers/infiniband/hw/i40iw/i40iw.h @@ -232,7 +232,7 @@ struct i40iw_device { struct i40e_client *client; struct i40iw_hw hw; struct i40iw_cm_core cm_core; - unsigned long *mem_resources; + u8 *mem_resources; unsigned long *allocated_qps; unsigned long *allocated_cqs; unsigned long *allocated_mrs; -- cgit From 44856be3e95c87f03e850ef4fdf8c0503c2dde18 Mon Sep 17 00:00:00 2001 From: Mustafa Ismail Date: Mon, 22 Aug 2016 18:15:58 -0500 Subject: i40iw: Protect req_resource_num update In i40iw_alloc_resource(), ensure that the update to req_resource_num is protected by the lock. Fixes: 8e06af711bf2 ("i40iw: add main, hdr, status") Reported-by: Stefan Assmann Signed-off-by: Mustafa Ismail Signed-off-by: Shiraz Saleem Signed-off-by: Doug Ledford --- drivers/infiniband/hw/i40iw/i40iw.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/i40iw/i40iw.h b/drivers/infiniband/hw/i40iw/i40iw.h index 882f3ef2e4c7..8ec09e470f84 100644 --- a/drivers/infiniband/hw/i40iw/i40iw.h +++ b/drivers/infiniband/hw/i40iw/i40iw.h @@ -435,8 +435,8 @@ static inline int i40iw_alloc_resource(struct i40iw_device *iwdev, *next = resource_num + 1; if (*next == max_resources) *next = 0; - spin_unlock_irqrestore(&iwdev->resource_lock, flags); *req_resource_num = resource_num; + spin_unlock_irqrestore(&iwdev->resource_lock, flags); return 0; } -- cgit From faa739fb5df56aadab96bcd2f6eb3486cc3a3aec Mon Sep 17 00:00:00 2001 From: Mustafa Ismail Date: Mon, 22 Aug 2016 18:17:12 -0500 Subject: i40iw: Add missing check for interface already open In i40iw_open(), check if interface is already open and return success if it is. Fixes: 8e06af711bf2 ("i40iw: add main, hdr, status") Reported-by: Stefan Assmann Signed-off-by: Mustafa Ismail Signed-off-by: Shiraz Saleem Signed-off-by: Doug Ledford --- drivers/infiniband/hw/i40iw/i40iw_main.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/i40iw/i40iw_main.c b/drivers/infiniband/hw/i40iw/i40iw_main.c index 6e9081380a27..0cbbe4038298 100644 --- a/drivers/infiniband/hw/i40iw/i40iw_main.c +++ b/drivers/infiniband/hw/i40iw/i40iw_main.c @@ -1558,6 +1558,10 @@ static int i40iw_open(struct i40e_info *ldev, struct i40e_client *client) enum i40iw_status_code status; struct i40iw_handler *hdl; + hdl = i40iw_find_netdev(ldev->netdev); + if (hdl) + return 0; + hdl = kzalloc(sizeof(*hdl), GFP_KERNEL); if (!hdl) return -ENOMEM; -- cgit From cff069b78c21559f427c3fefe9ef3294e3dec094 Mon Sep 17 00:00:00 2001 From: Bharat Potnuri Date: Tue, 23 Aug 2016 20:27:33 +0530 Subject: iw_cxgb4: Fix cxgb4 arm CQ logic w/IB_CQ_REPORT_MISSED_EVENTS Current cxgb4 arm CQ logic ignores IB_CQ_REPORT_MISSED_EVENTS for request completion notification on a CQ. Due to this ib_poll_handler() assumes all events polled and avoids further iopoll scheduling. This patch adds logic to cxgb4 ib_req_notify_cq() handler to check if CQ is not empty and return accordingly. Based on the return value of ib_req_notify_cq() handler, ib_poll_handler() will schedule a run of iopoll handler. Signed-off-by: Potnuri Bharat Teja Reviewed-by: Steve Wise Signed-off-by: Doug Ledford --- drivers/infiniband/hw/cxgb4/cq.c | 10 +++++----- drivers/infiniband/hw/cxgb4/t4.h | 5 +++++ 2 files changed, 10 insertions(+), 5 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/cxgb4/cq.c b/drivers/infiniband/hw/cxgb4/cq.c index 812ab7278b8e..ac926c942fee 100644 --- a/drivers/infiniband/hw/cxgb4/cq.c +++ b/drivers/infiniband/hw/cxgb4/cq.c @@ -1016,15 +1016,15 @@ int c4iw_resize_cq(struct ib_cq *cq, int cqe, struct ib_udata *udata) int c4iw_arm_cq(struct ib_cq *ibcq, enum ib_cq_notify_flags flags) { struct c4iw_cq *chp; - int ret; + int ret = 0; unsigned long flag; chp = to_c4iw_cq(ibcq); spin_lock_irqsave(&chp->lock, flag); - ret = t4_arm_cq(&chp->cq, - (flags & IB_CQ_SOLICITED_MASK) == IB_CQ_SOLICITED); + t4_arm_cq(&chp->cq, + (flags & IB_CQ_SOLICITED_MASK) == IB_CQ_SOLICITED); + if (flags & IB_CQ_REPORT_MISSED_EVENTS) + ret = t4_cq_notempty(&chp->cq); spin_unlock_irqrestore(&chp->lock, flag); - if (ret && !(flags & IB_CQ_REPORT_MISSED_EVENTS)) - ret = 0; return ret; } diff --git a/drivers/infiniband/hw/cxgb4/t4.h b/drivers/infiniband/hw/cxgb4/t4.h index 6126bbe36095..02173f4315fa 100644 --- a/drivers/infiniband/hw/cxgb4/t4.h +++ b/drivers/infiniband/hw/cxgb4/t4.h @@ -634,6 +634,11 @@ static inline int t4_valid_cqe(struct t4_cq *cq, struct t4_cqe *cqe) return (CQE_GENBIT(cqe) == cq->gen); } +static inline int t4_cq_notempty(struct t4_cq *cq) +{ + return cq->sw_in_use || t4_valid_cqe(cq, &cq->queue[cq->cidx]); +} + static inline int t4_next_hw_cqe(struct t4_cq *cq, struct t4_cqe **cqe) { int ret; -- cgit From 5dfd5e5e3bc68ab3912acc712c8180942094fc69 Mon Sep 17 00:00:00 2001 From: Shiraz Saleem Date: Mon, 22 Aug 2016 18:16:37 -0500 Subject: i40iw: Add missing NULL check for MPA private data Add NULL check for pdata and pdata->addr before the memcpy in i40iw_form_cm_frame(). This fixes a NULL pointer de-reference which occurs when the MPA private data pointer is NULL. Also only copy pdata->size bytes in the memcpy to prevent reading past the length of the private data buffer provided by upper layer. Fixes: f27b4746f378 ("i40iw: add connection management code") Reported-by: Stefan Assmann Signed-off-by: Mustafa Ismail Signed-off-by: Shiraz Saleem Signed-off-by: Doug Ledford --- drivers/infiniband/hw/i40iw/i40iw_cm.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/i40iw/i40iw_cm.c b/drivers/infiniband/hw/i40iw/i40iw_cm.c index 5026dc79978a..64343985e0b7 100644 --- a/drivers/infiniband/hw/i40iw/i40iw_cm.c +++ b/drivers/infiniband/hw/i40iw/i40iw_cm.c @@ -535,8 +535,8 @@ static struct i40iw_puda_buf *i40iw_form_cm_frame(struct i40iw_cm_node *cm_node, buf += hdr_len; } - if (pd_len) - memcpy(buf, pdata->addr, pd_len); + if (pdata && pdata->addr) + memcpy(buf, pdata->addr, pdata->size); atomic_set(&sqbuf->refcount, 1); -- cgit From 7eaf8313b1cfe93417a22bdc3f7380cac2a3dc6d Mon Sep 17 00:00:00 2001 From: Mustafa Ismail Date: Mon, 22 Aug 2016 19:01:47 -0500 Subject: i40iw: Do not set self-referencing pointer to NULL after kfree In i40iw_free_virt_mem(), do not set mem->va to NULL after freeing it as mem->va is a self-referencing pointer to mem. Fixes: 4e9042e647ff ("i40iw: add hw and utils files") Reported-by: Stefan Assmann Signed-off-by: Mustafa Ismail Signed-off-by: Shiraz Saleem Signed-off-by: Doug Ledford --- drivers/infiniband/hw/i40iw/i40iw_utils.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/i40iw/i40iw_utils.c b/drivers/infiniband/hw/i40iw/i40iw_utils.c index 0e8db0a35141..6fd043b1d714 100644 --- a/drivers/infiniband/hw/i40iw/i40iw_utils.c +++ b/drivers/infiniband/hw/i40iw/i40iw_utils.c @@ -673,8 +673,11 @@ enum i40iw_status_code i40iw_free_virt_mem(struct i40iw_hw *hw, { if (!mem) return I40IW_ERR_PARAM; + /* + * mem->va points to the parent of mem, so both mem and mem->va + * can not be touched once mem->va is freed + */ kfree(mem->va); - mem->va = NULL; return 0; } -- cgit From 82d200cc6fc817dbf049b67e5e3215eab427e846 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Tue, 23 Aug 2016 21:16:26 +0100 Subject: IB/mlx5: Remove superfluous include of io-mapping.h This file does not use any structs or functions defined by io-mapping.h (nor does it directly use iomap, ioremap, iounamp or friends). Remove it to simplify verification of changes to io-mapping.h The include existed since its inception in commit e126ba97dba9edeb6fafa3665b5f8497fc9cdf8c Author: Eli Cohen Date: Sun Jul 7 17:25:49 2013 +0300 mlx5: Add driver for Mellanox Connect-IB adapters which looks like a copy across from the Mellanox ethernet driver. Signed-off-by: Chris Wilson Cc: Eli Cohen Cc: Jack Morgenstein Cc: Or Gerlitz Cc: Matan Barak Cc: Leon Romanovsky Cc: Doug Ledford Cc: Sean Hefty Cc: Hal Rosenstock Cc: linux-rdma@vger.kernel.org Reviewed-by: Leon Romanovsky Reviewed-by: Laurence Oberman Tested-by: Laurence Oberman Signed-off-by: Doug Ledford --- drivers/infiniband/hw/mlx5/main.c | 1 - 1 file changed, 1 deletion(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c index a84bb766fc62..1b4094baa2de 100644 --- a/drivers/infiniband/hw/mlx5/main.c +++ b/drivers/infiniband/hw/mlx5/main.c @@ -37,7 +37,6 @@ #include #include #include -#include #if defined(CONFIG_X86) #include #endif -- cgit From d41d0910d97f05be987d2d60de7e8685c108963b Mon Sep 17 00:00:00 2001 From: Mustafa Ismail Date: Tue, 23 Aug 2016 16:50:13 -0500 Subject: i40iw: Fix double free of allocated_buffer Memory allocated for iwqp; iwqp->allocated_buffer is freed twice in the create_qp error path. Correct this by having it freed only once in i40iw_free_qp_resources(). Fixes: d37498417947 ("i40iw: add files for iwarp interface") Signed-off-by: Mustafa Ismail Signed-off-by: Shiraz Saleem Signed-off-by: Doug Ledford --- drivers/infiniband/hw/i40iw/i40iw_verbs.c | 1 - 1 file changed, 1 deletion(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/i40iw/i40iw_verbs.c b/drivers/infiniband/hw/i40iw/i40iw_verbs.c index 2360338877bf..722e5af1e3d8 100644 --- a/drivers/infiniband/hw/i40iw/i40iw_verbs.c +++ b/drivers/infiniband/hw/i40iw/i40iw_verbs.c @@ -794,7 +794,6 @@ static struct ib_qp *i40iw_create_qp(struct ib_pd *ibpd, return &iwqp->ibqp; error: i40iw_free_qp_resources(iwdev, iwqp, qp_num); - kfree(mem); return ERR_PTR(err_code); } -- cgit From 433c58139f6a7d59824aadd23d6c9cac1d4e6100 Mon Sep 17 00:00:00 2001 From: Mustafa Ismail Date: Tue, 23 Aug 2016 17:24:56 -0500 Subject: i40iw: Avoid writing to freed memory iwpbl->iwmr points to the structure that contains iwpbl, which is iwmr. Setting this to NULL would result in writing to freed memory. So just free iwmr, and return. Fixes: d37498417947 ("i40iw: add files for iwarp interface") Reported-by: Stefan Assmann Signed-off-by: Mustafa Ismail Signed-off-by: Shiraz Saleem Signed-off-by: Doug Ledford --- drivers/infiniband/hw/i40iw/i40iw_verbs.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/i40iw/i40iw_verbs.c b/drivers/infiniband/hw/i40iw/i40iw_verbs.c index 722e5af1e3d8..6329c971c22f 100644 --- a/drivers/infiniband/hw/i40iw/i40iw_verbs.c +++ b/drivers/infiniband/hw/i40iw/i40iw_verbs.c @@ -1925,8 +1925,7 @@ static int i40iw_dereg_mr(struct ib_mr *ib_mr) } if (iwpbl->pbl_allocated) i40iw_free_pble(iwdev->pble_rsrc, palloc); - kfree(iwpbl->iwmr); - iwpbl->iwmr = NULL; + kfree(iwmr); return 0; } -- cgit From 3c199b4523c92dc5df027eeeffa657e2ccf453ab Mon Sep 17 00:00:00 2001 From: Selvin Xavier Date: Wed, 24 Aug 2016 01:17:41 -0400 Subject: RDMA/ocrdma: Fix the max_sge reported from FW Current driver is reporting wrong values for max_sge and max_sge_rd in query_device. This breaks the nfs rdma and iser in some device profiles. Fixing the driver to report correct values from FW. Signed-off-by: Selvin Xavier Signed-off-by: Devesh Sharma Signed-off-by: Doug Ledford --- drivers/infiniband/hw/ocrdma/ocrdma_hw.c | 14 +++++++------- drivers/infiniband/hw/ocrdma/ocrdma_sli.h | 12 ++++++++---- drivers/infiniband/hw/ocrdma/ocrdma_verbs.c | 4 ++-- 3 files changed, 17 insertions(+), 13 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_hw.c b/drivers/infiniband/hw/ocrdma/ocrdma_hw.c index 16740dcb876b..67fc0b6857e1 100644 --- a/drivers/infiniband/hw/ocrdma/ocrdma_hw.c +++ b/drivers/infiniband/hw/ocrdma/ocrdma_hw.c @@ -1156,18 +1156,18 @@ static void ocrdma_get_attr(struct ocrdma_dev *dev, attr->max_srq = (rsp->max_srq_rpir_qps & OCRDMA_MBX_QUERY_CFG_MAX_SRQ_MASK) >> OCRDMA_MBX_QUERY_CFG_MAX_SRQ_OFFSET; - attr->max_send_sge = ((rsp->max_write_send_sge & + attr->max_send_sge = ((rsp->max_recv_send_sge & OCRDMA_MBX_QUERY_CFG_MAX_SEND_SGE_MASK) >> OCRDMA_MBX_QUERY_CFG_MAX_SEND_SGE_SHIFT); - attr->max_recv_sge = (rsp->max_write_send_sge & - OCRDMA_MBX_QUERY_CFG_MAX_SEND_SGE_MASK) >> - OCRDMA_MBX_QUERY_CFG_MAX_SEND_SGE_SHIFT; + attr->max_recv_sge = (rsp->max_recv_send_sge & + OCRDMA_MBX_QUERY_CFG_MAX_RECV_SGE_MASK) >> + OCRDMA_MBX_QUERY_CFG_MAX_RECV_SGE_SHIFT; attr->max_srq_sge = (rsp->max_srq_rqe_sge & OCRDMA_MBX_QUERY_CFG_MAX_SRQ_SGE_MASK) >> OCRDMA_MBX_QUERY_CFG_MAX_SRQ_SGE_OFFSET; - attr->max_rdma_sge = (rsp->max_write_send_sge & - OCRDMA_MBX_QUERY_CFG_MAX_WRITE_SGE_MASK) >> - OCRDMA_MBX_QUERY_CFG_MAX_WRITE_SGE_SHIFT; + attr->max_rdma_sge = (rsp->max_wr_rd_sge & + OCRDMA_MBX_QUERY_CFG_MAX_RD_SGE_MASK) >> + OCRDMA_MBX_QUERY_CFG_MAX_RD_SGE_SHIFT; attr->max_ord_per_qp = (rsp->max_ird_ord_per_qp & OCRDMA_MBX_QUERY_CFG_MAX_ORD_PER_QP_MASK) >> OCRDMA_MBX_QUERY_CFG_MAX_ORD_PER_QP_SHIFT; diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_sli.h b/drivers/infiniband/hw/ocrdma/ocrdma_sli.h index 0efc9662c6d8..37df4481bb8f 100644 --- a/drivers/infiniband/hw/ocrdma/ocrdma_sli.h +++ b/drivers/infiniband/hw/ocrdma/ocrdma_sli.h @@ -554,9 +554,9 @@ enum { OCRDMA_MBX_QUERY_CFG_L3_TYPE_MASK = 0x18, OCRDMA_MBX_QUERY_CFG_MAX_SEND_SGE_SHIFT = 0, OCRDMA_MBX_QUERY_CFG_MAX_SEND_SGE_MASK = 0xFFFF, - OCRDMA_MBX_QUERY_CFG_MAX_WRITE_SGE_SHIFT = 16, - OCRDMA_MBX_QUERY_CFG_MAX_WRITE_SGE_MASK = 0xFFFF << - OCRDMA_MBX_QUERY_CFG_MAX_WRITE_SGE_SHIFT, + OCRDMA_MBX_QUERY_CFG_MAX_RECV_SGE_SHIFT = 16, + OCRDMA_MBX_QUERY_CFG_MAX_RECV_SGE_MASK = 0xFFFF << + OCRDMA_MBX_QUERY_CFG_MAX_RECV_SGE_SHIFT, OCRDMA_MBX_QUERY_CFG_MAX_ORD_PER_QP_SHIFT = 0, OCRDMA_MBX_QUERY_CFG_MAX_ORD_PER_QP_MASK = 0xFFFF, @@ -612,6 +612,8 @@ enum { OCRDMA_MBX_QUERY_CFG_MAX_SRQ_SGE_OFFSET = 0, OCRDMA_MBX_QUERY_CFG_MAX_SRQ_SGE_MASK = 0xFFFF << OCRDMA_MBX_QUERY_CFG_MAX_SRQ_SGE_OFFSET, + OCRDMA_MBX_QUERY_CFG_MAX_RD_SGE_SHIFT = 0, + OCRDMA_MBX_QUERY_CFG_MAX_RD_SGE_MASK = 0xFFFF, }; struct ocrdma_mbx_query_config { @@ -619,7 +621,7 @@ struct ocrdma_mbx_query_config { struct ocrdma_mbx_rsp rsp; u32 qp_srq_cq_ird_ord; u32 max_pd_ca_ack_delay; - u32 max_write_send_sge; + u32 max_recv_send_sge; u32 max_ird_ord_per_qp; u32 max_shared_ird_ord; u32 max_mr; @@ -639,6 +641,8 @@ struct ocrdma_mbx_query_config { u32 max_wqes_rqes_per_q; u32 max_cq_cqes_per_cq; u32 max_srq_rqe_sge; + u32 max_wr_rd_sge; + u32 ird_pgsz_num_pages; }; struct ocrdma_fw_ver_rsp { diff --git a/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c b/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c index b1a3d91fe8b9..0aa854737e74 100644 --- a/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c +++ b/drivers/infiniband/hw/ocrdma/ocrdma_verbs.c @@ -125,8 +125,8 @@ int ocrdma_query_device(struct ib_device *ibdev, struct ib_device_attr *attr, IB_DEVICE_SYS_IMAGE_GUID | IB_DEVICE_LOCAL_DMA_LKEY | IB_DEVICE_MEM_MGT_EXTENSIONS; - attr->max_sge = dev->attr.max_send_sge; - attr->max_sge_rd = attr->max_sge; + attr->max_sge = min(dev->attr.max_send_sge, dev->attr.max_recv_sge); + attr->max_sge_rd = dev->attr.max_rdma_sge; attr->max_cq = dev->attr.max_cq; attr->max_cqe = dev->attr.max_cqe; attr->max_mr = dev->attr.max_mr; -- cgit From 07c72d7d54d138eb2ca37709a5a3d55fbcc01536 Mon Sep 17 00:00:00 2001 From: Tatyana Nikolova Date: Wed, 24 Aug 2016 13:59:17 -0500 Subject: i40iw: Send last streaming mode message for loopback connections Send a zero length last streaming mode message for loopback connections to synchronize between accepting QP and connecting QP. This avoids data transfer to start on the accepting QP before the connecting QP is in RTS. Also remove function i40iw_loopback_nop() as it is no longer used. Fixes: f27b4746f378 ("i40iw: add connection management code") Signed-off-by: Tatyana Nikolova Signed-off-by: Shiraz Saleem Signed-off-by: Doug Ledford --- drivers/infiniband/hw/i40iw/i40iw_cm.c | 22 +--------------------- 1 file changed, 1 insertion(+), 21 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/i40iw/i40iw_cm.c b/drivers/infiniband/hw/i40iw/i40iw_cm.c index 64343985e0b7..7ca0638579c0 100644 --- a/drivers/infiniband/hw/i40iw/i40iw_cm.c +++ b/drivers/infiniband/hw/i40iw/i40iw_cm.c @@ -3346,26 +3346,6 @@ int i40iw_cm_disconn(struct i40iw_qp *iwqp) return 0; } -/** - * i40iw_loopback_nop - Send a nop - * @qp: associated hw qp - */ -static void i40iw_loopback_nop(struct i40iw_sc_qp *qp) -{ - u64 *wqe; - u64 header; - - wqe = qp->qp_uk.sq_base->elem; - set_64bit_val(wqe, 0, 0); - set_64bit_val(wqe, 8, 0); - set_64bit_val(wqe, 16, 0); - - header = LS_64(I40IWQP_OP_NOP, I40IWQPSQ_OPCODE) | - LS_64(0, I40IWQPSQ_SIGCOMPL) | - LS_64(qp->qp_uk.swqe_polarity, I40IWQPSQ_VALID); - set_64bit_val(wqe, 24, header); -} - /** * i40iw_qp_disconnect - free qp and close cm * @iwqp: associate qp for the connection @@ -3638,7 +3618,7 @@ int i40iw_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) } else { if (iwqp->page) iwqp->sc_qp.qp_uk.sq_base = kmap(iwqp->page); - i40iw_loopback_nop(&iwqp->sc_qp); + dev->iw_priv_qp_ops->qp_send_lsmm(&iwqp->sc_qp, NULL, 0, 0); } if (iwqp->page) -- cgit From 866e0f4d73390ee6f5cd68aa92cf74eef3a2b0f2 Mon Sep 17 00:00:00 2001 From: Mustafa Ismail Date: Thu, 25 Aug 2016 11:52:47 -0500 Subject: i40iw: Update hw_iwarp_state Update iwqp->hw_iwarp_state to reflect the new state of the CQP modify QP operation. This avoids reissuing a CQP operation to modify a QP to a state that it is already in. Fixes: 4e9042e647ff ("i40iw: add hw and utils files") Reported-by: Stefan Assmann Signed-off-by: Mustafa Ismail Signed-off-by: Shiraz Saleem Signed-off-by: Doug Ledford --- drivers/infiniband/hw/i40iw/i40iw_hw.c | 1 + 1 file changed, 1 insertion(+) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/i40iw/i40iw_hw.c b/drivers/infiniband/hw/i40iw/i40iw_hw.c index 3ee0cad96bc6..0c92a40b3e86 100644 --- a/drivers/infiniband/hw/i40iw/i40iw_hw.c +++ b/drivers/infiniband/hw/i40iw/i40iw_hw.c @@ -265,6 +265,7 @@ void i40iw_next_iw_state(struct i40iw_qp *iwqp, info.dont_send_fin = false; if (iwqp->sc_qp.term_flags && (state == I40IW_QP_STATE_ERROR)) info.reset_tcp_conn = true; + iwqp->hw_iwarp_state = state; i40iw_hw_modify_qp(iwqp->iwdev, iwqp, &info, 0); } -- cgit From b71121b4b70a995c0b794026e84c880c4f26c361 Mon Sep 17 00:00:00 2001 From: Shiraz Saleem Date: Thu, 25 Aug 2016 11:53:24 -0500 Subject: i40iw: Receive notification events correctly Device notifications are not received after the first interface is closed; since there is an unregister for notifications on every interface close. Correct this by unregistering for device notifications only when the last interface is closed. Also, make all operations on the i40iw_notifiers_registered atomic as it can be read/modified concurrently. Fixes: 8e06af711bf2 ("i40iw: add main, hdr, status") Signed-off-by: Mustafa Ismail Signed-off-by: Shiraz Saleem Signed-off-by: Doug Ledford --- drivers/infiniband/hw/i40iw/i40iw_main.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/i40iw/i40iw_main.c b/drivers/infiniband/hw/i40iw/i40iw_main.c index 0cbbe4038298..445e230d5ff8 100644 --- a/drivers/infiniband/hw/i40iw/i40iw_main.c +++ b/drivers/infiniband/hw/i40iw/i40iw_main.c @@ -100,7 +100,7 @@ static struct notifier_block i40iw_net_notifier = { .notifier_call = i40iw_net_event }; -static int i40iw_notifiers_registered; +static atomic_t i40iw_notifiers_registered; /** * i40iw_find_i40e_handler - find a handler given a client info @@ -1342,12 +1342,11 @@ exit: */ static void i40iw_register_notifiers(void) { - if (!i40iw_notifiers_registered) { + if (atomic_inc_return(&i40iw_notifiers_registered) == 1) { register_inetaddr_notifier(&i40iw_inetaddr_notifier); register_inet6addr_notifier(&i40iw_inetaddr6_notifier); register_netevent_notifier(&i40iw_net_notifier); } - i40iw_notifiers_registered++; } /** @@ -1429,8 +1428,7 @@ static void i40iw_deinit_device(struct i40iw_device *iwdev, bool reset, bool del i40iw_del_macip_entry(iwdev, (u8)iwdev->mac_ip_table_idx); /* fallthrough */ case INET_NOTIFIER: - if (i40iw_notifiers_registered > 0) { - i40iw_notifiers_registered--; + if (!atomic_dec_return(&i40iw_notifiers_registered)) { unregister_netevent_notifier(&i40iw_net_notifier); unregister_inetaddr_notifier(&i40iw_inetaddr_notifier); unregister_inet6addr_notifier(&i40iw_inetaddr6_notifier); -- cgit From 61a28d2b690a8b6c383a075a12d62d711850f9d7 Mon Sep 17 00:00:00 2001 From: Christophe Jaillet Date: Fri, 26 Aug 2016 06:49:25 +0200 Subject: IB/hfi1: Clean up type used and casting In all other places in this file where 'find_first_bit' is called, port_num is defined as a 'u8' and no casting is done. Do the same here in order to be more consistent. Signed-off-by: Christophe JAILLET Signed-off-by: Doug Ledford --- drivers/infiniband/hw/hfi1/mad.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/hfi1/mad.c b/drivers/infiniband/hw/hfi1/mad.c index 39e42c373a01..9912d2ca3ded 100644 --- a/drivers/infiniband/hw/hfi1/mad.c +++ b/drivers/infiniband/hw/hfi1/mad.c @@ -2604,7 +2604,7 @@ static int pma_get_opa_datacounters(struct opa_pma_mad *pmp, u8 lq, num_vls; u8 res_lli, res_ler; u64 port_mask; - unsigned long port_num; + u8 port_num; unsigned long vl; u32 vl_select_mask; int vfi; @@ -2640,7 +2640,7 @@ static int pma_get_opa_datacounters(struct opa_pma_mad *pmp, port_num = find_first_bit((unsigned long *)&port_mask, sizeof(port_mask)); - if ((u8)port_num != port) { + if (port_num != port) { pmp->mad_hdr.status |= IB_SMP_INVALID_FIELD; return reply((struct ib_mad_hdr *)pmp); } -- cgit From fffd68734dc685e208e86d8c5f6522cd695a8d60 Mon Sep 17 00:00:00 2001 From: Christophe Jaillet Date: Fri, 26 Aug 2016 07:16:17 +0200 Subject: IB/mlx5: Fix the size parameter to find_first_bit The 2nd parameter of 'find_first_bit' is the number of bits to search. In this case, we are passing 'sizeof(tmp)' which is likely to be 4 or 8 because 'tmp' is an 'unsigned long'. It is likely that the number of bits of 'tmp' was expected here. So use BITS_PER_LONG instead. It has been spotted by the following coccinelle script: @@ expression ret, x; @@ * ret = \(find_first_bit \| find_first_zero_bit\) (x, sizeof(...)); Signed-off-by: Christophe JAILLET Acked-by: Majd Dibbiny Acked-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/hw/mlx5/mem.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/mlx5/mem.c b/drivers/infiniband/hw/mlx5/mem.c index 40df2cca0609..996b54e366b0 100644 --- a/drivers/infiniband/hw/mlx5/mem.c +++ b/drivers/infiniband/hw/mlx5/mem.c @@ -71,7 +71,7 @@ void mlx5_ib_cont_pages(struct ib_umem *umem, u64 addr, int *count, int *shift, addr = addr >> page_shift; tmp = (unsigned long)addr; - m = find_first_bit(&tmp, sizeof(tmp)); + m = find_first_bit(&tmp, BITS_PER_LONG); skip = 1 << m; mask = skip - 1; i = 0; @@ -81,7 +81,7 @@ void mlx5_ib_cont_pages(struct ib_umem *umem, u64 addr, int *count, int *shift, for (k = 0; k < len; k++) { if (!(i & mask)) { tmp = (unsigned long)pfn; - m = min_t(unsigned long, m, find_first_bit(&tmp, sizeof(tmp))); + m = min_t(unsigned long, m, find_first_bit(&tmp, BITS_PER_LONG)); skip = 1 << m; mask = skip - 1; base = pfn; @@ -89,7 +89,7 @@ void mlx5_ib_cont_pages(struct ib_umem *umem, u64 addr, int *count, int *shift, } else { if (base + p != pfn) { tmp = (unsigned long)p; - m = find_first_bit(&tmp, sizeof(tmp)); + m = find_first_bit(&tmp, BITS_PER_LONG); skip = 1 << m; mask = skip - 1; base = pfn; -- cgit From 6aaa382f1267644072f288916476879684502f73 Mon Sep 17 00:00:00 2001 From: Christophe Jaillet Date: Fri, 26 Aug 2016 06:49:09 +0200 Subject: IB/hfi1: Fix the size parameter to find_first_bit The 2nd parameter of 'find_first_bit' is the number of bits to search. In this case, we are passing 'sizeof(u64)' which is 8. It is likely that the number of bits of 'port_mask' was expected here. Use sizeof() * 8 to get the correct number. It has been spotted by the following coccinelle script: @@ expression ret, x; @@ * ret = \(find_first_bit \| find_first_zero_bit\) (x, sizeof(...)); Signed-off-by: Christophe JAILLET Signed-off-by: Doug Ledford --- drivers/infiniband/hw/hfi1/mad.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/hfi1/mad.c b/drivers/infiniband/hw/hfi1/mad.c index 9912d2ca3ded..7ffc14f21523 100644 --- a/drivers/infiniband/hw/hfi1/mad.c +++ b/drivers/infiniband/hw/hfi1/mad.c @@ -2638,7 +2638,7 @@ static int pma_get_opa_datacounters(struct opa_pma_mad *pmp, */ port_mask = be64_to_cpu(req->port_select_mask[3]); port_num = find_first_bit((unsigned long *)&port_mask, - sizeof(port_mask)); + sizeof(port_mask) * 8); if (port_num != port) { pmp->mad_hdr.status |= IB_SMP_INVALID_FIELD; @@ -2842,7 +2842,7 @@ static int pma_get_opa_porterrors(struct opa_pma_mad *pmp, */ port_mask = be64_to_cpu(req->port_select_mask[3]); port_num = find_first_bit((unsigned long *)&port_mask, - sizeof(port_mask)); + sizeof(port_mask) * 8); if (port_num != port) { pmp->mad_hdr.status |= IB_SMP_INVALID_FIELD; @@ -3015,7 +3015,7 @@ static int pma_get_opa_errorinfo(struct opa_pma_mad *pmp, */ port_mask = be64_to_cpu(req->port_select_mask[3]); port_num = find_first_bit((unsigned long *)&port_mask, - sizeof(port_mask)); + sizeof(port_mask) * 8); if (port_num != port) { pmp->mad_hdr.status |= IB_SMP_INVALID_FIELD; @@ -3252,7 +3252,7 @@ static int pma_set_opa_errorinfo(struct opa_pma_mad *pmp, */ port_mask = be64_to_cpu(req->port_select_mask[3]); port_num = find_first_bit((unsigned long *)&port_mask, - sizeof(port_mask)); + sizeof(port_mask) * 8); if (port_num != port) { pmp->mad_hdr.status |= IB_SMP_INVALID_FIELD; -- cgit From 656aacea6c90ce8e15c2bdef4f89b74b73e2e34a Mon Sep 17 00:00:00 2001 From: Baoyou Xie Date: Sun, 28 Aug 2016 22:57:11 +0800 Subject: IB/cxgb4: Make _free_qp static to silence build warning We get 1 warning when build kernel with W=1: drivers/infiniband/hw/cxgb4/qp.c:686:6: warning: no previous prototype for '_free_qp' [-Wmissing-prototypes] In fact, this function is only used in the file in which it is declared and don't need a declaration, but can be made static. so this patch marks it 'static'. Signed-off-by: Baoyou Xie Reviewed-by: Yuval Shaia Reviewed-by: Leon Romanovsky Acked-by: Steve Wise Signed-off-by: Doug Ledford --- drivers/infiniband/hw/cxgb4/qp.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/cxgb4/qp.c b/drivers/infiniband/hw/cxgb4/qp.c index edb1172b6f54..690435229be7 100644 --- a/drivers/infiniband/hw/cxgb4/qp.c +++ b/drivers/infiniband/hw/cxgb4/qp.c @@ -683,7 +683,7 @@ static int build_inv_stag(union t4_wr *wqe, struct ib_send_wr *wr, return 0; } -void _free_qp(struct kref *kref) +static void _free_qp(struct kref *kref) { struct c4iw_qp *qhp; -- cgit From 25b64fc5f2c0779b53ec155d69cc06b7cdc6e99f Mon Sep 17 00:00:00 2001 From: Leon Romanovsky Date: Sun, 28 Aug 2016 10:58:32 +0300 Subject: Revert "IB/mlx4: Return EAGAIN for any error in mlx4_ib_poll_one" By Mellanox HW design and SW implementation, poll_cq never fails and returns errors, so all these printks are to catch ULP bugs. In case of such bug, the reverted patch will cause reentry of the function, resulting in a printk storm. This reverts commit 5412352fcd8f ("IB/mlx4: Return EAGAIN for any error in mlx4_ib_poll_one") Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/hw/mlx4/cq.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/mlx4/cq.c b/drivers/infiniband/hw/mlx4/cq.c index 006db6436e3b..15b628996633 100644 --- a/drivers/infiniband/hw/mlx4/cq.c +++ b/drivers/infiniband/hw/mlx4/cq.c @@ -690,7 +690,7 @@ repoll: if (unlikely((cqe->owner_sr_opcode & MLX4_CQE_OPCODE_MASK) == MLX4_OPCODE_NOP && is_send)) { pr_warn("Completion for NOP opcode detected!\n"); - return -EAGAIN; + return -EINVAL; } /* Resize CQ in progress */ @@ -721,7 +721,7 @@ repoll: if (unlikely(!mqp)) { pr_warn("CQ %06x with entry for unknown QPN %06x\n", cq->mcq.cqn, be32_to_cpu(cqe->vlan_my_qpn) & MLX4_CQE_QPN_MASK); - return -EAGAIN; + return -EINVAL; } *cur_qp = to_mibqp(mqp); @@ -739,7 +739,7 @@ repoll: if (unlikely(!msrq)) { pr_warn("CQ %06x with entry for unknown SRQN %06x\n", cq->mcq.cqn, srq_num); - return -EAGAIN; + return -EINVAL; } } -- cgit From 20697434b6ea9c6d895ebc5217a46f18850a109f Mon Sep 17 00:00:00 2001 From: Leon Romanovsky Date: Sun, 28 Aug 2016 10:58:33 +0300 Subject: IB/mlx4: Don't return errors from poll_cq Remove returning errors from mlx4 poll_cq function. Polling CQ operation in kernel never fails by Mellanox HCA architecture and respective driver design. Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/hw/mlx4/cq.c | 26 ++------------------------ 1 file changed, 2 insertions(+), 24 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/mlx4/cq.c b/drivers/infiniband/hw/mlx4/cq.c index 15b628996633..5df63dacaaa3 100644 --- a/drivers/infiniband/hw/mlx4/cq.c +++ b/drivers/infiniband/hw/mlx4/cq.c @@ -687,12 +687,6 @@ repoll: is_error = (cqe->owner_sr_opcode & MLX4_CQE_OPCODE_MASK) == MLX4_CQE_OPCODE_ERROR; - if (unlikely((cqe->owner_sr_opcode & MLX4_CQE_OPCODE_MASK) == MLX4_OPCODE_NOP && - is_send)) { - pr_warn("Completion for NOP opcode detected!\n"); - return -EINVAL; - } - /* Resize CQ in progress */ if (unlikely((cqe->owner_sr_opcode & MLX4_CQE_OPCODE_MASK) == MLX4_CQE_OPCODE_RESIZE)) { if (cq->resize_buf) { @@ -718,12 +712,6 @@ repoll: */ mqp = __mlx4_qp_lookup(to_mdev(cq->ibcq.device)->dev, be32_to_cpu(cqe->vlan_my_qpn)); - if (unlikely(!mqp)) { - pr_warn("CQ %06x with entry for unknown QPN %06x\n", - cq->mcq.cqn, be32_to_cpu(cqe->vlan_my_qpn) & MLX4_CQE_QPN_MASK); - return -EINVAL; - } - *cur_qp = to_mibqp(mqp); } @@ -736,11 +724,6 @@ repoll: /* SRQ is also in the radix tree */ msrq = mlx4_srq_lookup(to_mdev(cq->ibcq.device)->dev, srq_num); - if (unlikely(!msrq)) { - pr_warn("CQ %06x with entry for unknown SRQN %06x\n", - cq->mcq.cqn, srq_num); - return -EINVAL; - } } if (is_send) { @@ -891,7 +874,6 @@ int mlx4_ib_poll_cq(struct ib_cq *ibcq, int num_entries, struct ib_wc *wc) struct mlx4_ib_qp *cur_qp = NULL; unsigned long flags; int npolled; - int err = 0; struct mlx4_ib_dev *mdev = to_mdev(cq->ibcq.device); spin_lock_irqsave(&cq->lock, flags); @@ -901,8 +883,7 @@ int mlx4_ib_poll_cq(struct ib_cq *ibcq, int num_entries, struct ib_wc *wc) } for (npolled = 0; npolled < num_entries; ++npolled) { - err = mlx4_ib_poll_one(cq, &cur_qp, wc + npolled); - if (err) + if (mlx4_ib_poll_one(cq, &cur_qp, wc + npolled)) break; } @@ -911,10 +892,7 @@ int mlx4_ib_poll_cq(struct ib_cq *ibcq, int num_entries, struct ib_wc *wc) out: spin_unlock_irqrestore(&cq->lock, flags); - if (err == 0 || err == -EAGAIN) - return npolled; - else - return err; + return npolled; } int mlx4_ib_arm_cq(struct ib_cq *ibcq, enum ib_cq_notify_flags flags) -- cgit From 24be409beede1a7cbe95b1740c4cdade3b6a8187 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Sun, 28 Aug 2016 10:58:34 +0300 Subject: IB/mlx5: Return EINVAL when caller specifies too many SGEs The returned value should be EINVAL, because it is caused by wrong caller and not by internal overflow event. Signed-off-by: Chuck Lever Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/hw/mlx5/qp.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/mlx5/qp.c b/drivers/infiniband/hw/mlx5/qp.c index 0dd7d93cac95..acb3b72e719e 100644 --- a/drivers/infiniband/hw/mlx5/qp.c +++ b/drivers/infiniband/hw/mlx5/qp.c @@ -3758,7 +3758,7 @@ int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr, num_sge = wr->num_sge; if (unlikely(num_sge > qp->sq.max_gs)) { mlx5_ib_warn(dev, "\n"); - err = -ENOMEM; + err = -EINVAL; *bad_wr = wr; goto out; } -- cgit From b2a232d21f301d600b02c6f9ccbc9f977331bb39 Mon Sep 17 00:00:00 2001 From: Leon Romanovsky Date: Sun, 28 Aug 2016 10:58:35 +0300 Subject: IB/mlx5: Simplify code by removing return variable Return variable was set in a line before the actual return was called in begin_wqe function. This patch removes such variable and simplifies the code. Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/hw/mlx5/qp.c | 10 +++------- 1 file changed, 3 insertions(+), 7 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/mlx5/qp.c b/drivers/infiniband/hw/mlx5/qp.c index acb3b72e719e..174d09b9728b 100644 --- a/drivers/infiniband/hw/mlx5/qp.c +++ b/drivers/infiniband/hw/mlx5/qp.c @@ -3658,12 +3658,8 @@ static int begin_wqe(struct mlx5_ib_qp *qp, void **seg, struct ib_send_wr *wr, unsigned *idx, int *size, int nreq) { - int err = 0; - - if (unlikely(mlx5_wq_overflow(&qp->sq, nreq, qp->ibqp.send_cq))) { - err = -ENOMEM; - return err; - } + if (unlikely(mlx5_wq_overflow(&qp->sq, nreq, qp->ibqp.send_cq))) + return -ENOMEM; *idx = qp->sq.cur_post & (qp->sq.wqe_cnt - 1); *seg = mlx5_get_send_wqe(qp, *idx); @@ -3679,7 +3675,7 @@ static int begin_wqe(struct mlx5_ib_qp *qp, void **seg, *seg += sizeof(**ctrl); *size = sizeof(**ctrl) / 16; - return err; + return 0; } static void finish_wqe(struct mlx5_ib_qp *qp, -- cgit From d9f88e5ab9a73058ebdde589219c0d37da250f06 Mon Sep 17 00:00:00 2001 From: Yishai Hadas Date: Sun, 28 Aug 2016 10:58:37 +0300 Subject: IB/mlx5: Use TIR number based on selector Use TIR number based on selector, it should be done to differentiate between RSS QP to RAW one. Reported-by: Sagi Grimberg Signed-off-by: Yishai Hadas Signed-off-by: Leon Romanovsky Tested-by: Sagi Grimberg Signed-off-by: Doug Ledford --- drivers/infiniband/hw/mlx5/main.c | 6 +++++- drivers/infiniband/hw/mlx5/mlx5_ib.h | 1 + drivers/infiniband/hw/mlx5/qp.c | 1 + 3 files changed, 7 insertions(+), 1 deletion(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c index 1b4094baa2de..8150ea372c53 100644 --- a/drivers/infiniband/hw/mlx5/main.c +++ b/drivers/infiniband/hw/mlx5/main.c @@ -1849,6 +1849,7 @@ static struct ib_flow *mlx5_ib_create_flow(struct ib_qp *qp, int domain) { struct mlx5_ib_dev *dev = to_mdev(qp->device); + struct mlx5_ib_qp *mqp = to_mqp(qp); struct mlx5_ib_flow_handler *handler = NULL; struct mlx5_flow_destination *dst = NULL; struct mlx5_ib_flow_prio *ft_prio; @@ -1875,7 +1876,10 @@ static struct ib_flow *mlx5_ib_create_flow(struct ib_qp *qp, } dst->type = MLX5_FLOW_DESTINATION_TYPE_TIR; - dst->tir_num = to_mqp(qp)->raw_packet_qp.rq.tirn; + if (mqp->flags & MLX5_IB_QP_RSS) + dst->tir_num = mqp->rss_qp.tirn; + else + dst->tir_num = mqp->raw_packet_qp.rq.tirn; if (flow_attr->type == IB_FLOW_ATTR_NORMAL) { if (flow_attr->flags & IB_FLOW_ATTR_FLAGS_DONT_TRAP) { diff --git a/drivers/infiniband/hw/mlx5/mlx5_ib.h b/drivers/infiniband/hw/mlx5/mlx5_ib.h index 372385d0f993..95146f4aa3e3 100644 --- a/drivers/infiniband/hw/mlx5/mlx5_ib.h +++ b/drivers/infiniband/hw/mlx5/mlx5_ib.h @@ -402,6 +402,7 @@ enum mlx5_ib_qp_flags { /* QP uses 1 as its source QP number */ MLX5_IB_QP_SQPN_QP1 = 1 << 6, MLX5_IB_QP_CAP_SCATTER_FCS = 1 << 7, + MLX5_IB_QP_RSS = 1 << 8, }; struct mlx5_umr_wr { diff --git a/drivers/infiniband/hw/mlx5/qp.c b/drivers/infiniband/hw/mlx5/qp.c index 174d09b9728b..affc3f6598ca 100644 --- a/drivers/infiniband/hw/mlx5/qp.c +++ b/drivers/infiniband/hw/mlx5/qp.c @@ -1449,6 +1449,7 @@ create_tir: kvfree(in); /* qpn is reserved for that QP */ qp->trans_qp.base.mqp.qpn = 0; + qp->flags |= MLX5_IB_QP_RSS; return 0; err: -- cgit From dbdf7d4e7f911f79ceb08365a756bbf6eecac81c Mon Sep 17 00:00:00 2001 From: Leon Romanovsky Date: Sun, 28 Aug 2016 10:58:38 +0300 Subject: IB/mlx5: Don't return errors from poll_cq Remove returning errors from mlx5 poll_cq function. Polling CQ operation in kernel never fails by Mellanox HCA architecture and respective driver design. Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/hw/mlx5/cq.c | 22 ++-------------------- 1 file changed, 2 insertions(+), 20 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/mlx5/cq.c b/drivers/infiniband/hw/mlx5/cq.c index 308a358e5b46..e4fac9292e4a 100644 --- a/drivers/infiniband/hw/mlx5/cq.c +++ b/drivers/infiniband/hw/mlx5/cq.c @@ -553,12 +553,6 @@ repoll: * from the table. */ mqp = __mlx5_qp_lookup(dev->mdev, qpn); - if (unlikely(!mqp)) { - mlx5_ib_warn(dev, "CQE@CQ %06x for unknown QPN %6x\n", - cq->mcq.cqn, qpn); - return -EINVAL; - } - *cur_qp = to_mibqp(mqp); } @@ -619,13 +613,6 @@ repoll: read_lock(&dev->mdev->priv.mkey_table.lock); mmkey = __mlx5_mr_lookup(dev->mdev, mlx5_base_mkey(be32_to_cpu(sig_err_cqe->mkey))); - if (unlikely(!mmkey)) { - read_unlock(&dev->mdev->priv.mkey_table.lock); - mlx5_ib_warn(dev, "CQE@CQ %06x for unknown MR %6x\n", - cq->mcq.cqn, be32_to_cpu(sig_err_cqe->mkey)); - return -EINVAL; - } - mr = to_mibmr(mmkey); get_sig_err_item(sig_err_cqe, &mr->sig->err_item); mr->sig->sig_err_exists = true; @@ -676,7 +663,6 @@ int mlx5_ib_poll_cq(struct ib_cq *ibcq, int num_entries, struct ib_wc *wc) unsigned long flags; int soft_polled = 0; int npolled; - int err = 0; spin_lock_irqsave(&cq->lock, flags); if (mdev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR) { @@ -688,8 +674,7 @@ int mlx5_ib_poll_cq(struct ib_cq *ibcq, int num_entries, struct ib_wc *wc) soft_polled = poll_soft_wc(cq, num_entries, wc); for (npolled = 0; npolled < num_entries - soft_polled; npolled++) { - err = mlx5_poll_one(cq, &cur_qp, wc + soft_polled + npolled); - if (err) + if (mlx5_poll_one(cq, &cur_qp, wc + soft_polled + npolled)) break; } @@ -698,10 +683,7 @@ int mlx5_ib_poll_cq(struct ib_cq *ibcq, int num_entries, struct ib_wc *wc) out: spin_unlock_irqrestore(&cq->lock, flags); - if (err == 0 || err == -EAGAIN) - return soft_polled + npolled; - else - return err; + return soft_polled + npolled; } int mlx5_ib_arm_cq(struct ib_cq *ibcq, enum ib_cq_notify_flags flags) -- cgit From 3e6c3b0fd5d071ed17bf91586aae35a6cfb8cdb3 Mon Sep 17 00:00:00 2001 From: Sebastian Sanchez Date: Wed, 31 Aug 2016 07:24:20 -0700 Subject: IB/hfi1: Fix SGE length for misaligned PIO copy When trying to align the source pointer and there's a byte carry in an SGE copy, bytes are borrowed from the next quad-word X to complete the required quad-word copy. Then, the SGE length is reduced by the number of borrowed bytes. After this, if the remaining number of bytes from quad-word X (extra bytes) is greater than the new SGE length, the number of extra bytes needs to be updated to the new SGE length. Otherwise, when the SGE length gets updated again after the extra bytes are read to create the new byte carry, it goes negative, which then becomes a very large number as the SGE length is an unsigned integer. This causes SGE buffer to be over-read. Reviewed-by: Dean Luick Signed-off-by: Sebastian Sanchez Signed-off-by: Dennis Dalessandro Signed-off-by: Doug Ledford --- drivers/infiniband/hw/hfi1/pio_copy.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/hfi1/pio_copy.c b/drivers/infiniband/hw/hfi1/pio_copy.c index 8c25e1b58849..3a1ef3056282 100644 --- a/drivers/infiniband/hw/hfi1/pio_copy.c +++ b/drivers/infiniband/hw/hfi1/pio_copy.c @@ -771,6 +771,9 @@ void seg_pio_copy_mid(struct pio_buf *pbuf, const void *from, size_t nbytes) read_extra_bytes(pbuf, from, to_fill); from += to_fill; nbytes -= to_fill; + /* may not be enough valid bytes left to align */ + if (extra > nbytes) + extra = nbytes; /* ...now write carry */ dest = pbuf->start + (pbuf->qw_written * sizeof(u64)); @@ -798,6 +801,15 @@ void seg_pio_copy_mid(struct pio_buf *pbuf, const void *from, size_t nbytes) read_low_bytes(pbuf, from, extra); from += extra; nbytes -= extra; + /* + * If no bytes are left, return early - we are done. + * NOTE: This short-circuit is *required* because + * "extra" may have been reduced in size and "from" + * is not aligned, as required when leaving this + * if block. + */ + if (nbytes == 0) + return; } /* at this point, from is QW aligned */ -- cgit From af53493916693343955930556aaa83f875b8436a Mon Sep 17 00:00:00 2001 From: Jubin John Date: Wed, 31 Aug 2016 07:24:27 -0700 Subject: IB/hfi1: Fix AHG KDETH Intr shift In the set_txreq_header_ahg(), The KDETH Intr bit is obtained from the header in the user sdma request using a KDETH_GET shift and mask macro. This value is then futher right shifted by 16 causing us to lose the value i.e it is shifted to zero, leading to the following smatch warning: drivers/infiniband/hw/hfi1/user_sdma.c:1482 set_txreq_header_ahg() warn: mask and shift to zero The Intr bit should be left shifted into its correct position in the KDETH header before the AHG update. Reported-by: Dan Carpenter Reviewed-by: Mitko Haralanov Reviewed-by: Harish Chegondi Signed-off-by: Jubin John Signed-off-by: Dennis Dalessandro Signed-off-by: Doug Ledford --- drivers/infiniband/hw/hfi1/user_sdma.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/hfi1/user_sdma.c b/drivers/infiniband/hw/hfi1/user_sdma.c index 0ecf27903dc2..1694037d1eee 100644 --- a/drivers/infiniband/hw/hfi1/user_sdma.c +++ b/drivers/infiniband/hw/hfi1/user_sdma.c @@ -114,6 +114,8 @@ MODULE_PARM_DESC(sdma_comp_size, "Size of User SDMA completion ring. Default: 12 #define KDETH_HCRC_LOWER_SHIFT 24 #define KDETH_HCRC_LOWER_MASK 0xff +#define AHG_KDETH_INTR_SHIFT 12 + #define PBC2LRH(x) ((((x) & 0xfff) << 2) - 4) #define LRH2PBC(x) ((((x) >> 2) + 1) & 0xfff) @@ -1480,7 +1482,8 @@ static int set_txreq_header_ahg(struct user_sdma_request *req, /* Clear KDETH.SH on last packet */ if (unlikely(tx->flags & TXREQ_FLAGS_REQ_LAST_PKT)) { val |= cpu_to_le16(KDETH_GET(hdr->kdeth.ver_tid_offset, - INTR) >> 16); + INTR) << + AHG_KDETH_INTR_SHIFT); val &= cpu_to_le16(~(1U << 13)); AHG_HEADER_SET(req->ahg, diff, 7, 16, 14, val); } else { -- cgit From 673b975f1fbad5382f76afdb987e7513c5f4b71b Mon Sep 17 00:00:00 2001 From: Dean Luick Date: Wed, 31 Aug 2016 07:24:33 -0700 Subject: IB/hfi1: Add QSFP sanity pre-check Sometimes a QSFP device does not respond in the expected time after a power-on. Add a read pre-check/retry when starting the link on driver load. Reviewed-by: Easwar Hariharan Signed-off-by: Dean Luick Signed-off-by: Dennis Dalessandro Signed-off-by: Doug Ledford --- drivers/infiniband/hw/hfi1/chip.c | 86 +++++++++++++++++++++++++++++++++++---- drivers/infiniband/hw/hfi1/chip.h | 1 + drivers/infiniband/hw/hfi1/hfi.h | 2 + drivers/infiniband/hw/hfi1/init.c | 1 + 4 files changed, 82 insertions(+), 8 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/hfi1/chip.c b/drivers/infiniband/hw/hfi1/chip.c index b32638d58ae8..ec3635a7ee3e 100644 --- a/drivers/infiniband/hw/hfi1/chip.c +++ b/drivers/infiniband/hw/hfi1/chip.c @@ -9490,6 +9490,78 @@ static void init_lcb(struct hfi1_devdata *dd) write_csr(dd, DC_LCB_CFG_TX_FIFOS_RESET, 0x00); } +/* + * Perform a test read on the QSFP. Return 0 on success, -ERRNO + * on error. + */ +static int test_qsfp_read(struct hfi1_pportdata *ppd) +{ + int ret; + u8 status; + + /* report success if not a QSFP */ + if (ppd->port_type != PORT_TYPE_QSFP) + return 0; + + /* read byte 2, the status byte */ + ret = one_qsfp_read(ppd, ppd->dd->hfi1_id, 2, &status, 1); + if (ret < 0) + return ret; + if (ret != 1) + return -EIO; + + return 0; /* success */ +} + +/* + * Values for QSFP retry. + * + * Give up after 10s (20 x 500ms). The overall timeout was empirically + * arrived at from experience on a large cluster. + */ +#define MAX_QSFP_RETRIES 20 +#define QSFP_RETRY_WAIT 500 /* msec */ + +/* + * Try a QSFP read. If it fails, schedule a retry for later. + * Called on first link activation after driver load. + */ +static void try_start_link(struct hfi1_pportdata *ppd) +{ + if (test_qsfp_read(ppd)) { + /* read failed */ + if (ppd->qsfp_retry_count >= MAX_QSFP_RETRIES) { + dd_dev_err(ppd->dd, "QSFP not responding, giving up\n"); + return; + } + dd_dev_info(ppd->dd, + "QSFP not responding, waiting and retrying %d\n", + (int)ppd->qsfp_retry_count); + ppd->qsfp_retry_count++; + queue_delayed_work(ppd->hfi1_wq, &ppd->start_link_work, + msecs_to_jiffies(QSFP_RETRY_WAIT)); + return; + } + ppd->qsfp_retry_count = 0; + + /* + * Tune the SerDes to a ballpark setting for optimal signal and bit + * error rate. Needs to be done before starting the link. + */ + tune_serdes(ppd); + start_link(ppd); +} + +/* + * Workqueue function to start the link after a delay. + */ +void handle_start_link(struct work_struct *work) +{ + struct hfi1_pportdata *ppd = container_of(work, struct hfi1_pportdata, + start_link_work.work); + try_start_link(ppd); +} + int bringup_serdes(struct hfi1_pportdata *ppd) { struct hfi1_devdata *dd = ppd->dd; @@ -9525,14 +9597,8 @@ int bringup_serdes(struct hfi1_pportdata *ppd) set_qsfp_int_n(ppd, 1); } - /* - * Tune the SerDes to a ballpark setting for - * optimal signal and bit error rate - * Needs to be done before starting the link - */ - tune_serdes(ppd); - - return start_link(ppd); + try_start_link(ppd); + return 0; } void hfi1_quiet_serdes(struct hfi1_pportdata *ppd) @@ -9549,6 +9615,10 @@ void hfi1_quiet_serdes(struct hfi1_pportdata *ppd) ppd->driver_link_ready = 0; ppd->link_enabled = 0; + ppd->qsfp_retry_count = MAX_QSFP_RETRIES; /* prevent more retries */ + flush_delayed_work(&ppd->start_link_work); + cancel_delayed_work_sync(&ppd->start_link_work); + ppd->offline_disabled_reason = HFI1_ODR_MASK(OPA_LINKDOWN_REASON_SMA_DISABLED); set_link_down_reason(ppd, OPA_LINKDOWN_REASON_SMA_DISABLED, 0, diff --git a/drivers/infiniband/hw/hfi1/chip.h b/drivers/infiniband/hw/hfi1/chip.h index ed11107c50fe..e29573769efc 100644 --- a/drivers/infiniband/hw/hfi1/chip.h +++ b/drivers/infiniband/hw/hfi1/chip.h @@ -706,6 +706,7 @@ void handle_link_up(struct work_struct *work); void handle_link_down(struct work_struct *work); void handle_link_downgrade(struct work_struct *work); void handle_link_bounce(struct work_struct *work); +void handle_start_link(struct work_struct *work); void handle_sma_message(struct work_struct *work); void reset_qsfp(struct hfi1_pportdata *ppd); void qsfp_event(struct work_struct *work); diff --git a/drivers/infiniband/hw/hfi1/hfi.h b/drivers/infiniband/hw/hfi1/hfi.h index a021e660d482..28b91286798b 100644 --- a/drivers/infiniband/hw/hfi1/hfi.h +++ b/drivers/infiniband/hw/hfi1/hfi.h @@ -605,6 +605,7 @@ struct hfi1_pportdata { struct work_struct freeze_work; struct work_struct link_downgrade_work; struct work_struct link_bounce_work; + struct delayed_work start_link_work; /* host link state variables */ struct mutex hls_lock; u32 host_link_state; @@ -659,6 +660,7 @@ struct hfi1_pportdata { u8 linkinit_reason; u8 local_tx_rate; /* rate given to 8051 firmware */ u8 last_pstate; /* info only */ + u8 qsfp_retry_count; /* placeholders for IB MAD packet settings */ u8 overrun_threshold; diff --git a/drivers/infiniband/hw/hfi1/init.c b/drivers/infiniband/hw/hfi1/init.c index b7935451093c..000dc079bd88 100644 --- a/drivers/infiniband/hw/hfi1/init.c +++ b/drivers/infiniband/hw/hfi1/init.c @@ -500,6 +500,7 @@ void hfi1_init_pportdata(struct pci_dev *pdev, struct hfi1_pportdata *ppd, INIT_WORK(&ppd->link_downgrade_work, handle_link_downgrade); INIT_WORK(&ppd->sma_message_work, handle_sma_message); INIT_WORK(&ppd->link_bounce_work, handle_link_bounce); + INIT_DELAYED_WORK(&ppd->start_link_work, handle_start_link); INIT_WORK(&ppd->linkstate_active_work, receive_interrupt_work); INIT_WORK(&ppd->qsfp_info.qsfp_work, qsfp_event); -- cgit From 429b6a721727d49d8565b50a6bc0dc42432383a9 Mon Sep 17 00:00:00 2001 From: Harish Chegondi Date: Wed, 31 Aug 2016 07:24:40 -0700 Subject: IB/hfi1: Make n_krcvqs be an unsigned long integer The global variable n_krcvqs stores the sum of the number of kernel receive queues of VLs 0-7 which the user can pass to the driver through the module parameter array krcvqs which is of type unsigned integer. If the user passes large value(s) into krcvqs parameter array, it can cause an arithmetic overflow while calculating n_krcvqs which is also of type unsigned int. The overflow results in an incorrect value of n_krcvqs which can lead to kernel crash while loading the driver. Fix by changing the data type of n_krcvqs to unsigned long. This patch also changes the data type of other variables that get their values from n_krcvqs. Reviewed-by: Dennis Dalessandro Signed-off-by: Harish Chegondi Signed-off-by: Dennis Dalessandro Signed-off-by: Doug Ledford --- drivers/infiniband/hw/hfi1/chip.c | 6 +++--- drivers/infiniband/hw/hfi1/hfi.h | 2 +- drivers/infiniband/hw/hfi1/init.c | 2 +- 3 files changed, 5 insertions(+), 5 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/hfi1/chip.c b/drivers/infiniband/hw/hfi1/chip.c index ec3635a7ee3e..cc38004cea42 100644 --- a/drivers/infiniband/hw/hfi1/chip.c +++ b/drivers/infiniband/hw/hfi1/chip.c @@ -12935,7 +12935,7 @@ fail: */ static int set_up_context_variables(struct hfi1_devdata *dd) { - int num_kernel_contexts; + unsigned long num_kernel_contexts; int total_contexts; int ret; unsigned ngroups; @@ -12964,9 +12964,9 @@ static int set_up_context_variables(struct hfi1_devdata *dd) */ if (num_kernel_contexts > (dd->chip_send_contexts - num_vls - 1)) { dd_dev_err(dd, - "Reducing # kernel rcv contexts to: %d, from %d\n", + "Reducing # kernel rcv contexts to: %d, from %lu\n", (int)(dd->chip_send_contexts - num_vls - 1), - (int)num_kernel_contexts); + num_kernel_contexts); num_kernel_contexts = dd->chip_send_contexts - num_vls - 1; } /* diff --git a/drivers/infiniband/hw/hfi1/hfi.h b/drivers/infiniband/hw/hfi1/hfi.h index 28b91286798b..325ec211370f 100644 --- a/drivers/infiniband/hw/hfi1/hfi.h +++ b/drivers/infiniband/hw/hfi1/hfi.h @@ -1806,7 +1806,7 @@ extern unsigned int hfi1_max_mtu; extern unsigned int hfi1_cu; extern unsigned int user_credit_return_threshold; extern int num_user_contexts; -extern unsigned n_krcvqs; +extern unsigned long n_krcvqs; extern uint krcvqs[]; extern int krcvqsset; extern uint kdeth_qp; diff --git a/drivers/infiniband/hw/hfi1/init.c b/drivers/infiniband/hw/hfi1/init.c index 000dc079bd88..384b43d2fd49 100644 --- a/drivers/infiniband/hw/hfi1/init.c +++ b/drivers/infiniband/hw/hfi1/init.c @@ -94,7 +94,7 @@ module_param_array(krcvqs, uint, &krcvqsset, S_IRUGO); MODULE_PARM_DESC(krcvqs, "Array of the number of non-control kernel receive queues by VL"); /* computed based on above array */ -unsigned n_krcvqs; +unsigned long n_krcvqs; static unsigned hfi1_rcvarr_split = 25; module_param_named(rcvarr_split, hfi1_rcvarr_split, uint, S_IRUGO); -- cgit From 16170d9c102764f76c58aad244e947f4e3f44590 Mon Sep 17 00:00:00 2001 From: Mike Marciniszyn Date: Wed, 31 Aug 2016 07:24:46 -0700 Subject: IB/hfi1: Rework debugfs to use SRCU The debugfs RCU trips many debug kernel warnings because of potential sleeps with an RCU read lock held. This includes both user copy calls and slab allocations throughout the file. This patch switches the RCU to use SRCU for file remove/access race protection. In one case, the SRCU is implicit in the use of the raw debugfs file object and just works. In the seq_file case, a wrapper around seq_read() and seq_lseek() is used to enforce the SRCU using the debugfs supplied functions debugfs_use_file_start() and debugfs_use_file_stop(). The sychronize_rcu() is deleted since the SRCU prevents the remove access race. The RCU locking is kept for qp_stats since the QP hash list is protected using the non-sleepable RCU. Reviewed-by: Sebastian Sanchez Signed-off-by: Mike Marciniszyn Signed-off-by: Dennis Dalessandro Signed-off-by: Doug Ledford --- drivers/infiniband/hw/hfi1/debugfs.c | 132 ++++++++++++++--------------------- 1 file changed, 52 insertions(+), 80 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/hfi1/debugfs.c b/drivers/infiniband/hw/hfi1/debugfs.c index a49cc88f08a2..5e9be16f6cd3 100644 --- a/drivers/infiniband/hw/hfi1/debugfs.c +++ b/drivers/infiniband/hw/hfi1/debugfs.c @@ -59,6 +59,40 @@ static struct dentry *hfi1_dbg_root; +/* wrappers to enforce srcu in seq file */ +static ssize_t hfi1_seq_read( + struct file *file, + char __user *buf, + size_t size, + loff_t *ppos) +{ + struct dentry *d = file->f_path.dentry; + int srcu_idx; + ssize_t r; + + r = debugfs_use_file_start(d, &srcu_idx); + if (likely(!r)) + r = seq_read(file, buf, size, ppos); + debugfs_use_file_finish(srcu_idx); + return r; +} + +static loff_t hfi1_seq_lseek( + struct file *file, + loff_t offset, + int whence) +{ + struct dentry *d = file->f_path.dentry; + int srcu_idx; + loff_t r; + + r = debugfs_use_file_start(d, &srcu_idx); + if (likely(!r)) + r = seq_lseek(file, offset, whence); + debugfs_use_file_finish(srcu_idx); + return r; +} + #define private2dd(file) (file_inode(file)->i_private) #define private2ppd(file) (file_inode(file)->i_private) @@ -87,8 +121,8 @@ static int _##name##_open(struct inode *inode, struct file *s) \ static const struct file_operations _##name##_file_ops = { \ .owner = THIS_MODULE, \ .open = _##name##_open, \ - .read = seq_read, \ - .llseek = seq_lseek, \ + .read = hfi1_seq_read, \ + .llseek = hfi1_seq_lseek, \ .release = seq_release \ } @@ -105,11 +139,9 @@ do { \ DEBUGFS_FILE_CREATE(#name, parent, data, &_##name##_file_ops, S_IRUGO) static void *_opcode_stats_seq_start(struct seq_file *s, loff_t *pos) -__acquires(RCU) { struct hfi1_opcode_stats_perctx *opstats; - rcu_read_lock(); if (*pos >= ARRAY_SIZE(opstats->stats)) return NULL; return pos; @@ -126,9 +158,7 @@ static void *_opcode_stats_seq_next(struct seq_file *s, void *v, loff_t *pos) } static void _opcode_stats_seq_stop(struct seq_file *s, void *v) -__releases(RCU) { - rcu_read_unlock(); } static int _opcode_stats_seq_show(struct seq_file *s, void *v) @@ -285,12 +315,10 @@ DEBUGFS_SEQ_FILE_OPEN(qp_stats) DEBUGFS_FILE_OPS(qp_stats); static void *_sdes_seq_start(struct seq_file *s, loff_t *pos) -__acquires(RCU) { struct hfi1_ibdev *ibd; struct hfi1_devdata *dd; - rcu_read_lock(); ibd = (struct hfi1_ibdev *)s->private; dd = dd_from_dev(ibd); if (!dd->per_sdma || *pos >= dd->num_sdma) @@ -310,9 +338,7 @@ static void *_sdes_seq_next(struct seq_file *s, void *v, loff_t *pos) } static void _sdes_seq_stop(struct seq_file *s, void *v) -__releases(RCU) { - rcu_read_unlock(); } static int _sdes_seq_show(struct seq_file *s, void *v) @@ -339,11 +365,9 @@ static ssize_t dev_counters_read(struct file *file, char __user *buf, struct hfi1_devdata *dd; ssize_t rval; - rcu_read_lock(); dd = private2dd(file); avail = hfi1_read_cntrs(dd, NULL, &counters); rval = simple_read_from_buffer(buf, count, ppos, counters, avail); - rcu_read_unlock(); return rval; } @@ -356,11 +380,9 @@ static ssize_t dev_names_read(struct file *file, char __user *buf, struct hfi1_devdata *dd; ssize_t rval; - rcu_read_lock(); dd = private2dd(file); avail = hfi1_read_cntrs(dd, &names, NULL); rval = simple_read_from_buffer(buf, count, ppos, names, avail); - rcu_read_unlock(); return rval; } @@ -383,11 +405,9 @@ static ssize_t portnames_read(struct file *file, char __user *buf, struct hfi1_devdata *dd; ssize_t rval; - rcu_read_lock(); dd = private2dd(file); avail = hfi1_read_portcntrs(dd->pport, &names, NULL); rval = simple_read_from_buffer(buf, count, ppos, names, avail); - rcu_read_unlock(); return rval; } @@ -400,11 +420,9 @@ static ssize_t portcntrs_debugfs_read(struct file *file, char __user *buf, struct hfi1_pportdata *ppd; ssize_t rval; - rcu_read_lock(); ppd = private2ppd(file); avail = hfi1_read_portcntrs(ppd, NULL, &counters); rval = simple_read_from_buffer(buf, count, ppos, counters, avail); - rcu_read_unlock(); return rval; } @@ -434,16 +452,13 @@ static ssize_t asic_flags_read(struct file *file, char __user *buf, int used; int i; - rcu_read_lock(); ppd = private2ppd(file); dd = ppd->dd; size = PAGE_SIZE; used = 0; tmp = kmalloc(size, GFP_KERNEL); - if (!tmp) { - rcu_read_unlock(); + if (!tmp) return -ENOMEM; - } scratch0 = read_csr(dd, ASIC_CFG_SCRATCH); used += scnprintf(tmp + used, size - used, @@ -470,7 +485,6 @@ static ssize_t asic_flags_read(struct file *file, char __user *buf, used += scnprintf(tmp + used, size - used, "Write bits to clear\n"); ret = simple_read_from_buffer(buf, count, ppos, tmp, used); - rcu_read_unlock(); kfree(tmp); return ret; } @@ -486,15 +500,12 @@ static ssize_t asic_flags_write(struct file *file, const char __user *buf, u64 scratch0; u64 clear; - rcu_read_lock(); ppd = private2ppd(file); dd = ppd->dd; buff = kmalloc(count + 1, GFP_KERNEL); - if (!buff) { - ret = -ENOMEM; - goto do_return; - } + if (!buff) + return -ENOMEM; ret = copy_from_user(buff, buf, count); if (ret > 0) { @@ -527,8 +538,6 @@ static ssize_t asic_flags_write(struct file *file, const char __user *buf, do_free: kfree(buff); - do_return: - rcu_read_unlock(); return ret; } @@ -542,18 +551,14 @@ static ssize_t qsfp_debugfs_dump(struct file *file, char __user *buf, char *tmp; int ret; - rcu_read_lock(); ppd = private2ppd(file); tmp = kmalloc(PAGE_SIZE, GFP_KERNEL); - if (!tmp) { - rcu_read_unlock(); + if (!tmp) return -ENOMEM; - } ret = qsfp_dump(ppd, tmp, PAGE_SIZE); if (ret > 0) ret = simple_read_from_buffer(buf, count, ppos, tmp, ret); - rcu_read_unlock(); kfree(tmp); return ret; } @@ -569,7 +574,6 @@ static ssize_t __i2c_debugfs_write(struct file *file, const char __user *buf, int offset; int total_written; - rcu_read_lock(); ppd = private2ppd(file); /* byte offset format: [offsetSize][i2cAddr][offsetHigh][offsetLow] */ @@ -577,16 +581,12 @@ static ssize_t __i2c_debugfs_write(struct file *file, const char __user *buf, offset = *ppos & 0xffff; /* explicitly reject invalid address 0 to catch cp and cat */ - if (i2c_addr == 0) { - ret = -EINVAL; - goto _return; - } + if (i2c_addr == 0) + return -EINVAL; buff = kmalloc(count, GFP_KERNEL); - if (!buff) { - ret = -ENOMEM; - goto _return; - } + if (!buff) + return -ENOMEM; ret = copy_from_user(buff, buf, count); if (ret > 0) { @@ -606,8 +606,6 @@ static ssize_t __i2c_debugfs_write(struct file *file, const char __user *buf, _free: kfree(buff); - _return: - rcu_read_unlock(); return ret; } @@ -636,7 +634,6 @@ static ssize_t __i2c_debugfs_read(struct file *file, char __user *buf, int offset; int total_read; - rcu_read_lock(); ppd = private2ppd(file); /* byte offset format: [offsetSize][i2cAddr][offsetHigh][offsetLow] */ @@ -644,16 +641,12 @@ static ssize_t __i2c_debugfs_read(struct file *file, char __user *buf, offset = *ppos & 0xffff; /* explicitly reject invalid address 0 to catch cp and cat */ - if (i2c_addr == 0) { - ret = -EINVAL; - goto _return; - } + if (i2c_addr == 0) + return -EINVAL; buff = kmalloc(count, GFP_KERNEL); - if (!buff) { - ret = -ENOMEM; - goto _return; - } + if (!buff) + return -ENOMEM; total_read = i2c_read(ppd, target, i2c_addr, offset, buff, count); if (total_read < 0) { @@ -673,8 +666,6 @@ static ssize_t __i2c_debugfs_read(struct file *file, char __user *buf, _free: kfree(buff); - _return: - rcu_read_unlock(); return ret; } @@ -701,26 +692,20 @@ static ssize_t __qsfp_debugfs_write(struct file *file, const char __user *buf, int ret; int total_written; - rcu_read_lock(); - if (*ppos + count > QSFP_PAGESIZE * 4) { /* base page + page00-page03 */ - ret = -EINVAL; - goto _return; - } + if (*ppos + count > QSFP_PAGESIZE * 4) /* base page + page00-page03 */ + return -EINVAL; ppd = private2ppd(file); buff = kmalloc(count, GFP_KERNEL); - if (!buff) { - ret = -ENOMEM; - goto _return; - } + if (!buff) + return -ENOMEM; ret = copy_from_user(buff, buf, count); if (ret > 0) { ret = -EFAULT; goto _free; } - total_written = qsfp_write(ppd, target, *ppos, buff, count); if (total_written < 0) { ret = total_written; @@ -733,8 +718,6 @@ static ssize_t __qsfp_debugfs_write(struct file *file, const char __user *buf, _free: kfree(buff); - _return: - rcu_read_unlock(); return ret; } @@ -761,7 +744,6 @@ static ssize_t __qsfp_debugfs_read(struct file *file, char __user *buf, int ret; int total_read; - rcu_read_lock(); if (*ppos + count > QSFP_PAGESIZE * 4) { /* base page + page00-page03 */ ret = -EINVAL; goto _return; @@ -794,7 +776,6 @@ static ssize_t __qsfp_debugfs_read(struct file *file, char __user *buf, _free: kfree(buff); _return: - rcu_read_unlock(); return ret; } @@ -1010,7 +991,6 @@ void hfi1_dbg_ibdev_exit(struct hfi1_ibdev *ibd) debugfs_remove_recursive(ibd->hfi1_ibdev_dbg); out: ibd->hfi1_ibdev_dbg = NULL; - synchronize_rcu(); } /* @@ -1035,9 +1015,7 @@ static const char * const hfi1_statnames[] = { }; static void *_driver_stats_names_seq_start(struct seq_file *s, loff_t *pos) -__acquires(RCU) { - rcu_read_lock(); if (*pos >= ARRAY_SIZE(hfi1_statnames)) return NULL; return pos; @@ -1055,9 +1033,7 @@ static void *_driver_stats_names_seq_next( } static void _driver_stats_names_seq_stop(struct seq_file *s, void *v) -__releases(RCU) { - rcu_read_unlock(); } static int _driver_stats_names_seq_show(struct seq_file *s, void *v) @@ -1073,9 +1049,7 @@ DEBUGFS_SEQ_FILE_OPEN(driver_stats_names) DEBUGFS_FILE_OPS(driver_stats_names); static void *_driver_stats_seq_start(struct seq_file *s, loff_t *pos) -__acquires(RCU) { - rcu_read_lock(); if (*pos >= ARRAY_SIZE(hfi1_statnames)) return NULL; return pos; @@ -1090,9 +1064,7 @@ static void *_driver_stats_seq_next(struct seq_file *s, void *v, loff_t *pos) } static void _driver_stats_seq_stop(struct seq_file *s, void *v) -__releases(RCU) { - rcu_read_unlock(); } static u64 hfi1_sps_ints(void) -- cgit From 609e941a6bcd7ceb1cbb561941c997f6465e8698 Mon Sep 17 00:00:00 2001 From: Steve Wise Date: Thu, 1 Sep 2016 06:43:46 -0700 Subject: iw_cxgb4: call dev_put() on l2t allocation failure Reviewed-by: Sagi Grimberg Reviewed-by: Christoph Hellwig Signed-off-by: Steve Wise Signed-off-by: Sagi Grimberg --- drivers/infiniband/hw/cxgb4/cm.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/cxgb4/cm.c b/drivers/infiniband/hw/cxgb4/cm.c index a3a67216bce6..57b5bb56cdbf 100644 --- a/drivers/infiniband/hw/cxgb4/cm.c +++ b/drivers/infiniband/hw/cxgb4/cm.c @@ -2077,8 +2077,10 @@ static int import_ep(struct c4iw_ep *ep, int iptype, __u8 *peer_ip, } ep->l2t = cxgb4_l2t_get(cdev->rdev.lldi.l2t, n, pdev, rt_tos2priority(tos)); - if (!ep->l2t) + if (!ep->l2t) { + dev_put(pdev); goto out; + } ep->mtu = pdev->mtu; ep->tx_chan = cxgb4_port_chan(pdev); ep->smac_idx = cxgb4_tp_smt_idx(adapter_type, -- cgit From 37eb816c0867b1b0db273d22b530780a0a083980 Mon Sep 17 00:00:00 2001 From: Steve Wise Date: Thu, 1 Sep 2016 06:44:52 -0700 Subject: iw_cxgb4: block module unload until all ep resources are released Otherwise an endpoint can be still closing down causing a touch after free crash. Also WARN_ON if ulps have failed to destroy various resources during device removal. Fixes: ad61a4c7a9b7 ("iw_cxgb4: don't block in destroy_qp awaiting the last deref") Reviewed-by: Sagi Grimberg Reviewed-by: Christoph Hellwig Signed-off-by: Steve Wise Signed-off-by: Sagi Grimberg --- drivers/infiniband/hw/cxgb4/cm.c | 2 ++ drivers/infiniband/hw/cxgb4/device.c | 5 +++++ drivers/infiniband/hw/cxgb4/iw_cxgb4.h | 1 + 3 files changed, 8 insertions(+) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/cxgb4/cm.c b/drivers/infiniband/hw/cxgb4/cm.c index 57b5bb56cdbf..5621270664f4 100644 --- a/drivers/infiniband/hw/cxgb4/cm.c +++ b/drivers/infiniband/hw/cxgb4/cm.c @@ -314,6 +314,8 @@ static void remove_ep_tid(struct c4iw_ep *ep) spin_lock_irqsave(&ep->com.dev->lock, flags); _remove_handle(ep->com.dev, &ep->com.dev->hwtid_idr, ep->hwtid, 0); + if (idr_is_empty(&ep->com.dev->hwtid_idr)) + wake_up(&ep->com.dev->wait); spin_unlock_irqrestore(&ep->com.dev->lock, flags); } diff --git a/drivers/infiniband/hw/cxgb4/device.c b/drivers/infiniband/hw/cxgb4/device.c index ae2e8b23d2dd..63561aa9c69e 100644 --- a/drivers/infiniband/hw/cxgb4/device.c +++ b/drivers/infiniband/hw/cxgb4/device.c @@ -872,9 +872,13 @@ static void c4iw_rdev_close(struct c4iw_rdev *rdev) static void c4iw_dealloc(struct uld_ctx *ctx) { c4iw_rdev_close(&ctx->dev->rdev); + WARN_ON_ONCE(!idr_is_empty(&ctx->dev->cqidr)); idr_destroy(&ctx->dev->cqidr); + WARN_ON_ONCE(!idr_is_empty(&ctx->dev->qpidr)); idr_destroy(&ctx->dev->qpidr); + WARN_ON_ONCE(!idr_is_empty(&ctx->dev->mmidr)); idr_destroy(&ctx->dev->mmidr); + wait_event(ctx->dev->wait, idr_is_empty(&ctx->dev->hwtid_idr)); idr_destroy(&ctx->dev->hwtid_idr); idr_destroy(&ctx->dev->stid_idr); idr_destroy(&ctx->dev->atid_idr); @@ -992,6 +996,7 @@ static struct c4iw_dev *c4iw_alloc(const struct cxgb4_lld_info *infop) mutex_init(&devp->rdev.stats.lock); mutex_init(&devp->db_mutex); INIT_LIST_HEAD(&devp->db_fc_list); + init_waitqueue_head(&devp->wait); devp->avail_ird = devp->rdev.lldi.max_ird_adapter; if (c4iw_debugfs_root) { diff --git a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h index f6f34a75af27..0e9cd4479bb3 100644 --- a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h +++ b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h @@ -263,6 +263,7 @@ struct c4iw_dev { struct idr stid_idr; struct list_head db_fc_list; u32 avail_ird; + wait_queue_head_t wait; }; static inline struct c4iw_dev *to_c4iw_dev(struct ib_device *ibdev) -- cgit From e5ac40cd66c2f3cd11bc5edc658f012661b16347 Mon Sep 17 00:00:00 2001 From: Alex Vesker Date: Mon, 12 Sep 2016 19:16:18 +0300 Subject: IB/mlx4: Fix incorrect MC join state bit-masking on SR-IOV Because of an incorrect bit-masking done on the join state bits, when handling a join request we failed to detect a difference between the group join state and the request join state when joining as send only full member (0x8). This caused the MC join request not to be sent. This issue is relevant only when SRIOV is enabled and SM supports send only full member. This fix separates scope bits and join states bits a nibble each. Fixes: b9c5d6a64358 ('IB/mlx4: Add multicast group (MCG) paravirtualization for SR-IOV') Signed-off-by: Alex Vesker Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/hw/mlx4/mcg.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/mlx4/mcg.c b/drivers/infiniband/hw/mlx4/mcg.c index 8f7ad07915b0..097bfcc4ee99 100644 --- a/drivers/infiniband/hw/mlx4/mcg.c +++ b/drivers/infiniband/hw/mlx4/mcg.c @@ -489,7 +489,7 @@ static u8 get_leave_state(struct mcast_group *group) if (!group->members[i]) leave_state |= (1 << i); - return leave_state & (group->rec.scope_join_state & 7); + return leave_state & (group->rec.scope_join_state & 0xf); } static int join_group(struct mcast_group *group, int slave, u8 join_mask) @@ -564,8 +564,8 @@ static void mlx4_ib_mcg_timeout_handler(struct work_struct *work) } else mcg_warn_group(group, "DRIVER BUG\n"); } else if (group->state == MCAST_LEAVE_SENT) { - if (group->rec.scope_join_state & 7) - group->rec.scope_join_state &= 0xf8; + if (group->rec.scope_join_state & 0xf) + group->rec.scope_join_state &= 0xf0; group->state = MCAST_IDLE; mutex_unlock(&group->lock); if (release_group(group, 1)) @@ -605,7 +605,7 @@ static int handle_leave_req(struct mcast_group *group, u8 leave_mask, static int handle_join_req(struct mcast_group *group, u8 join_mask, struct mcast_req *req) { - u8 group_join_state = group->rec.scope_join_state & 7; + u8 group_join_state = group->rec.scope_join_state & 0xf; int ref = 0; u16 status; struct ib_sa_mcmember_data *sa_data = (struct ib_sa_mcmember_data *)req->sa_mad.data; @@ -690,8 +690,8 @@ static void mlx4_ib_mcg_work_handler(struct work_struct *work) u8 cur_join_state; resp_join_state = ((struct ib_sa_mcmember_data *) - group->response_sa_mad.data)->scope_join_state & 7; - cur_join_state = group->rec.scope_join_state & 7; + group->response_sa_mad.data)->scope_join_state & 0xf; + cur_join_state = group->rec.scope_join_state & 0xf; if (method == IB_MGMT_METHOD_GET_RESP) { /* successfull join */ @@ -710,7 +710,7 @@ process_requests: req = list_first_entry(&group->pending_list, struct mcast_req, group_list); sa_data = (struct ib_sa_mcmember_data *)req->sa_mad.data; - req_join_state = sa_data->scope_join_state & 0x7; + req_join_state = sa_data->scope_join_state & 0xf; /* For a leave request, we will immediately answer the VF, and * update our internal counters. The actual leave will be sent -- cgit From baa0be7026e2f7d1d40bfd45909044169e9e3c68 Mon Sep 17 00:00:00 2001 From: Jack Morgenstein Date: Mon, 12 Sep 2016 19:16:19 +0300 Subject: IB/mlx4: Fix code indentation in QP1 MAD flow The indentation in the QP1 GRH flow in procedure build_mlx_header is really confusing. Fix it, in preparation for a commit which touches this code. Fixes: 1ffeb2eb8be9 ("IB/mlx4: SR-IOV IB context objects and proxy/tunnel SQP support") Signed-off-by: Jack Morgenstein Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/hw/mlx4/qp.c | 36 +++++++++++++++++++----------------- 1 file changed, 19 insertions(+), 17 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/mlx4/qp.c b/drivers/infiniband/hw/mlx4/qp.c index 768085f59566..e398c04903fa 100644 --- a/drivers/infiniband/hw/mlx4/qp.c +++ b/drivers/infiniband/hw/mlx4/qp.c @@ -2493,24 +2493,26 @@ static int build_mlx_header(struct mlx4_ib_sqp *sqp, struct ib_ud_wr *wr, sqp->ud_header.grh.flow_label = ah->av.ib.sl_tclass_flowlabel & cpu_to_be32(0xfffff); sqp->ud_header.grh.hop_limit = ah->av.ib.hop_limit; - if (is_eth) + if (is_eth) { memcpy(sqp->ud_header.grh.source_gid.raw, sgid.raw, 16); - else { - if (mlx4_is_mfunc(to_mdev(ib_dev)->dev)) { - /* When multi-function is enabled, the ib_core gid - * indexes don't necessarily match the hw ones, so - * we must use our own cache */ - sqp->ud_header.grh.source_gid.global.subnet_prefix = - to_mdev(ib_dev)->sriov.demux[sqp->qp.port - 1]. - subnet_prefix; - sqp->ud_header.grh.source_gid.global.interface_id = - to_mdev(ib_dev)->sriov.demux[sqp->qp.port - 1]. - guid_cache[ah->av.ib.gid_index]; - } else - ib_get_cached_gid(ib_dev, - be32_to_cpu(ah->av.ib.port_pd) >> 24, - ah->av.ib.gid_index, - &sqp->ud_header.grh.source_gid, NULL); + } else { + if (mlx4_is_mfunc(to_mdev(ib_dev)->dev)) { + /* When multi-function is enabled, the ib_core gid + * indexes don't necessarily match the hw ones, so + * we must use our own cache + */ + sqp->ud_header.grh.source_gid.global.subnet_prefix = + to_mdev(ib_dev)->sriov.demux[sqp->qp.port - 1]. + subnet_prefix; + sqp->ud_header.grh.source_gid.global.interface_id = + to_mdev(ib_dev)->sriov.demux[sqp->qp.port - 1]. + guid_cache[ah->av.ib.gid_index]; + } else { + ib_get_cached_gid(ib_dev, + be32_to_cpu(ah->av.ib.port_pd) >> 24, + ah->av.ib.gid_index, + &sqp->ud_header.grh.source_gid, NULL); + } } memcpy(sqp->ud_header.grh.destination_gid.raw, ah->av.ib.dgid, 16); -- cgit From 8ec07bf8a8b57d6c58927a16a0a22c0115cf2855 Mon Sep 17 00:00:00 2001 From: Jack Morgenstein Date: Mon, 12 Sep 2016 19:16:20 +0300 Subject: IB/mlx4: Use correct subnet-prefix in QP1 mads under SR-IOV When sending QP1 MAD packets which use a GRH, the source GID (which consists of the 64-bit subnet prefix, and the 64 bit port GUID) must be included in the packet GRH. For SR-IOV, a GID cache is used, since the source GID needs to be the slave's source GID, and not the Hypervisor's GID. This cache also included a subnet_prefix. Unfortunately, the subnet_prefix field in the cache was never initialized (to the default subnet prefix 0xfe80::0). As a result, this field remained all zeroes. Therefore, when SR-IOV was active, all QP1 packets which included a GRH had a source GID subnet prefix of all-zeroes. However, the subnet-prefix should initially be 0xfe80::0 (the default subnet prefix). In addition, if OpenSM modifies a port's subnet prefix, the new subnet prefix must be used in the GRH when sending QP1 packets. To fix this we now initialize the subnet prefix in the SR-IOV GID cache to the default subnet prefix. We update the cached value if/when OpenSM modifies the port's subnet prefix. We take this cached value when sending QP1 packets when SR-IOV is active. Note that the value is stored as an atomic64. This eliminates any need for locking when the subnet prefix is being updated. Note also that we depend on the FW generating the "port management change" event for tracking subnet-prefix changes performed by OpenSM. If running early FW (before 2.9.4630), subnet prefix changes will not be tracked (but the default subnet prefix still will be stored in the cache; therefore users who do not modify the subnet prefix will not have a problem). IF there is a need for such tracking also for early FW, we will add that capability in a subsequent patch. Fixes: 1ffeb2eb8be9 ("IB/mlx4: SR-IOV IB context objects and proxy/tunnel SQP support") Signed-off-by: Jack Morgenstein Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/hw/mlx4/mad.c | 23 +++++++++++++++++++++++ drivers/infiniband/hw/mlx4/mlx4_ib.h | 2 +- drivers/infiniband/hw/mlx4/qp.c | 5 +++-- 3 files changed, 27 insertions(+), 3 deletions(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/mlx4/mad.c b/drivers/infiniband/hw/mlx4/mad.c index 9c2e53d28f98..0f21c3a25552 100644 --- a/drivers/infiniband/hw/mlx4/mad.c +++ b/drivers/infiniband/hw/mlx4/mad.c @@ -1128,6 +1128,27 @@ void handle_port_mgmt_change_event(struct work_struct *work) /* Generate GUID changed event */ if (changed_attr & MLX4_EQ_PORT_INFO_GID_PFX_CHANGE_MASK) { + if (mlx4_is_master(dev->dev)) { + union ib_gid gid; + int err = 0; + + if (!eqe->event.port_mgmt_change.params.port_info.gid_prefix) + err = __mlx4_ib_query_gid(&dev->ib_dev, port, 0, &gid, 1); + else + gid.global.subnet_prefix = + eqe->event.port_mgmt_change.params.port_info.gid_prefix; + if (err) { + pr_warn("Could not change QP1 subnet prefix for port %d: query_gid error (%d)\n", + port, err); + } else { + pr_debug("Changing QP1 subnet prefix for port %d. old=0x%llx. new=0x%llx\n", + port, + (u64)atomic64_read(&dev->sriov.demux[port - 1].subnet_prefix), + be64_to_cpu(gid.global.subnet_prefix)); + atomic64_set(&dev->sriov.demux[port - 1].subnet_prefix, + be64_to_cpu(gid.global.subnet_prefix)); + } + } mlx4_ib_dispatch_event(dev, port, IB_EVENT_GID_CHANGE); /*if master, notify all slaves*/ if (mlx4_is_master(dev->dev)) @@ -2202,6 +2223,8 @@ int mlx4_ib_init_sriov(struct mlx4_ib_dev *dev) if (err) goto demux_err; dev->sriov.demux[i].guid_cache[0] = gid.global.interface_id; + atomic64_set(&dev->sriov.demux[i].subnet_prefix, + be64_to_cpu(gid.global.subnet_prefix)); err = alloc_pv_object(dev, mlx4_master_func_num(dev->dev), i + 1, &dev->sriov.sqps[i]); if (err) diff --git a/drivers/infiniband/hw/mlx4/mlx4_ib.h b/drivers/infiniband/hw/mlx4/mlx4_ib.h index 7c5832ede4bd..686ab48ff644 100644 --- a/drivers/infiniband/hw/mlx4/mlx4_ib.h +++ b/drivers/infiniband/hw/mlx4/mlx4_ib.h @@ -448,7 +448,7 @@ struct mlx4_ib_demux_ctx { struct workqueue_struct *wq; struct workqueue_struct *ud_wq; spinlock_t ud_lock; - __be64 subnet_prefix; + atomic64_t subnet_prefix; __be64 guid_cache[128]; struct mlx4_ib_dev *dev; /* the following lock protects both mcg_table and mcg_mgid0_list */ diff --git a/drivers/infiniband/hw/mlx4/qp.c b/drivers/infiniband/hw/mlx4/qp.c index e398c04903fa..7fb9629bd12b 100644 --- a/drivers/infiniband/hw/mlx4/qp.c +++ b/drivers/infiniband/hw/mlx4/qp.c @@ -2502,8 +2502,9 @@ static int build_mlx_header(struct mlx4_ib_sqp *sqp, struct ib_ud_wr *wr, * we must use our own cache */ sqp->ud_header.grh.source_gid.global.subnet_prefix = - to_mdev(ib_dev)->sriov.demux[sqp->qp.port - 1]. - subnet_prefix; + cpu_to_be64(atomic64_read(&(to_mdev(ib_dev)->sriov. + demux[sqp->qp.port - 1]. + subnet_prefix))); sqp->ud_header.grh.source_gid.global.interface_id = to_mdev(ib_dev)->sriov.demux[sqp->qp.port - 1]. guid_cache[ah->av.ib.gid_index]; -- cgit From 69d269d38910e697e461ec5677368f57d2046cbe Mon Sep 17 00:00:00 2001 From: Kamal Heib Date: Mon, 12 Sep 2016 19:16:22 +0300 Subject: IB/mlx4: Diagnostic HW counters are not supported in slave mode Modify the mlx4_ib_diag_counters() to avoid the following error in the hypervisor when the slave tries to query the hardware counters in SR-IOV mode. mlx4_core 0000:81:00.0: Unknown command:0x30 accepted from slave:1 Fixes: 3f85f2aaabf7 ("IB/mlx4: Add diagnostic hardware counters") Signed-off-by: Kamal Heib Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/hw/mlx4/main.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c index 2af44c2de262..87ba9bca4181 100644 --- a/drivers/infiniband/hw/mlx4/main.c +++ b/drivers/infiniband/hw/mlx4/main.c @@ -2202,6 +2202,9 @@ static int mlx4_ib_alloc_diag_counters(struct mlx4_ib_dev *ibdev) bool per_port = !!(ibdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_DIAG_PER_PORT); + if (mlx4_is_slave(ibdev->dev)) + return 0; + for (i = 0; i < MLX4_DIAG_COUNTERS_TYPES; i++) { /* i == 1 means we are building port counters */ if (i && !per_port) -- cgit From 7fae6655a0c897875bd34501ec092232b526d3e4 Mon Sep 17 00:00:00 2001 From: Noa Osherovich Date: Mon, 12 Sep 2016 19:16:23 +0300 Subject: IB/mlx5: Enable MAD_IFC commands for IB ports only MAD_IFC command is supported only for physical functions (PF) and when physical port is IB. The proposed fix enforces it. Fixes: d603c809ef91 ("IB/mlx5: Fix decision on using MAD_IFC") Reported-by: David Chang Signed-off-by: Noa Osherovich Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/hw/mlx5/main.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c index 8150ea372c53..0480b643fb9c 100644 --- a/drivers/infiniband/hw/mlx5/main.c +++ b/drivers/infiniband/hw/mlx5/main.c @@ -288,7 +288,9 @@ __be16 mlx5_get_roce_udp_sport(struct mlx5_ib_dev *dev, u8 port_num, static int mlx5_use_mad_ifc(struct mlx5_ib_dev *dev) { - return !MLX5_CAP_GEN(dev->mdev, ib_virt); + if (MLX5_CAP_GEN(dev->mdev, port_type) == MLX5_CAP_PORT_TYPE_IB) + return !MLX5_CAP_GEN(dev->mdev, ib_virt); + return 0; } enum { -- cgit From ee3da804ad1b1dd4c766199a6e8443542b0aaaef Mon Sep 17 00:00:00 2001 From: Maor Gottlieb Date: Mon, 12 Sep 2016 19:16:24 +0300 Subject: IB/mlx5: Set source mac address in FTE Set the source mac address in the FTE when L2 specification is provided. Fixes: 038d2ef87572 ('IB/mlx5: Add flow steering support') Signed-off-by: Maor Gottlieb Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/hw/mlx5/main.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'drivers/infiniband/hw') diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c index 0480b643fb9c..e19537cf44ab 100644 --- a/drivers/infiniband/hw/mlx5/main.c +++ b/drivers/infiniband/hw/mlx5/main.c @@ -1430,6 +1430,13 @@ static int parse_flow_attr(u32 *match_c, u32 *match_v, dmac_47_16), ib_spec->eth.val.dst_mac); + ether_addr_copy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, outer_headers_c, + smac_47_16), + ib_spec->eth.mask.src_mac); + ether_addr_copy(MLX5_ADDR_OF(fte_match_set_lyr_2_4, outer_headers_v, + smac_47_16), + ib_spec->eth.val.src_mac); + if (ib_spec->eth.mask.vlan_tag) { MLX5_SET(fte_match_set_lyr_2_4, outer_headers_c, vlan_tag, 1); -- cgit