From f3e48119b97f56fb09310c95d49da122a27003d7 Mon Sep 17 00:00:00 2001 From: Ram Amrani Date: Tue, 14 Mar 2017 15:25:58 +0200 Subject: [PATCH 1/7] qed: Align CIDs according to DORQ requirement The Doorbell HW block can be configured at a granularity of 16 x CIDs, so we need to make sure that the actual number of CIDs configured would be a multiplication of 16. Today, when RoCE is enabled - given that the number is unaligned, doorbelling the higher CIDs would fail to reach the firmware and would eventually timeout. Fixes: dbb799c39717 ("qed: Initialize hardware for new protocols") Signed-off-by: Ram Amrani Signed-off-by: Yuval Mintz Signed-off-by: David S. Miller --- drivers/net/ethernet/qlogic/qed/qed_cxt.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/qlogic/qed/qed_cxt.c b/drivers/net/ethernet/qlogic/qed/qed_cxt.c index d42d03df751acb..7e3a6fed3da6d9 100644 --- a/drivers/net/ethernet/qlogic/qed/qed_cxt.c +++ b/drivers/net/ethernet/qlogic/qed/qed_cxt.c @@ -422,8 +422,9 @@ static void qed_cxt_set_proto_cid_count(struct qed_hwfn *p_hwfn, u32 page_sz = p_mgr->clients[ILT_CLI_CDUC].p_size.val; u32 cxt_size = CONN_CXT_SIZE(p_hwfn); u32 elems_per_page = ILT_PAGE_IN_BYTES(page_sz) / cxt_size; + u32 align = elems_per_page * DQ_RANGE_ALIGN; - p_conn->cid_count = roundup(p_conn->cid_count, elems_per_page); + p_conn->cid_count = roundup(p_conn->cid_count, align); } } From 3ef310a7d99216e0fbdff29f0cb13bc54180373a Mon Sep 17 00:00:00 2001 From: Tomer Tayar Date: Tue, 14 Mar 2017 15:25:59 +0200 Subject: [PATCH 2/7] qed: Prevent creation of too-big u32-chains Current Logic would allow the creation of a chain with U32_MAX + 1 elements, when the actual maximum supported by the driver infrastructure is U32_MAX. Fixes: a91eb52abb50 ("qed: Revisit chain implementation") Signed-off-by: Tomer Tayar Signed-off-by: Yuval Mintz Signed-off-by: David S. Miller --- drivers/net/ethernet/qlogic/qed/qed_dev.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/drivers/net/ethernet/qlogic/qed/qed_dev.c b/drivers/net/ethernet/qlogic/qed/qed_dev.c index e2a081ceaf520c..e518f914eab13f 100644 --- a/drivers/net/ethernet/qlogic/qed/qed_dev.c +++ b/drivers/net/ethernet/qlogic/qed/qed_dev.c @@ -2389,9 +2389,8 @@ qed_chain_alloc_sanity_check(struct qed_dev *cdev, * size/capacity fields are of a u32 type. */ if ((cnt_type == QED_CHAIN_CNT_TYPE_U16 && - chain_size > 0x10000) || - (cnt_type == QED_CHAIN_CNT_TYPE_U32 && - chain_size > 0x100000000ULL)) { + chain_size > ((u32)U16_MAX + 1)) || + (cnt_type == QED_CHAIN_CNT_TYPE_U32 && chain_size > U32_MAX)) { DP_NOTICE(cdev, "The actual chain size (0x%llx) is larger than the maximal possible value\n", chain_size); From 752ecb2da11124a948567076b60767dc8034cfa5 Mon Sep 17 00:00:00 2001 From: "Mintz, Yuval" Date: Tue, 14 Mar 2017 15:26:00 +0200 Subject: [PATCH 3/7] qed: Fix mapping leak on LL2 rx flow When receiving an Rx LL2 packet, qed fails to unmap the previous buffer. Fixes: 0a7fb11c23c0 ("qed: Add Light L2 support"); Signed-off-by: Yuval Mintz Signed-off-by: David S. Miller --- drivers/net/ethernet/qlogic/qed/qed_ll2.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/net/ethernet/qlogic/qed/qed_ll2.c b/drivers/net/ethernet/qlogic/qed/qed_ll2.c index 5fb34db377c8c3..29ae5ec2ac2db6 100644 --- a/drivers/net/ethernet/qlogic/qed/qed_ll2.c +++ b/drivers/net/ethernet/qlogic/qed/qed_ll2.c @@ -211,6 +211,8 @@ static void qed_ll2b_complete_rx_packet(struct qed_hwfn *p_hwfn, /* If need to reuse or there's no replacement buffer, repost this */ if (rc) goto out_post; + dma_unmap_single(&cdev->pdev->dev, buffer->phys_addr, + cdev->ll2->rx_size, DMA_FROM_DEVICE); skb = build_skb(buffer->data, 0); if (!skb) { From 4621ceb279d065151eb940ce8a4728b10c0646c7 Mon Sep 17 00:00:00 2001 From: "Mintz, Yuval" Date: Tue, 14 Mar 2017 15:26:01 +0200 Subject: [PATCH 4/7] qed: Free previous connections when releasing iSCSI Fixes: fc831825f99e ("qed: Add support for hardware offloaded iSCSI") Signed-off-by: Yuval Mintz Signed-off-by: David S. Miller --- drivers/net/ethernet/qlogic/qed/qed_iscsi.c | 28 +++++++++++++++++++++ 1 file changed, 28 insertions(+) diff --git a/drivers/net/ethernet/qlogic/qed/qed_iscsi.c b/drivers/net/ethernet/qlogic/qed/qed_iscsi.c index 3a44d6b395fac9..7e73b05eeab862 100644 --- a/drivers/net/ethernet/qlogic/qed/qed_iscsi.c +++ b/drivers/net/ethernet/qlogic/qed/qed_iscsi.c @@ -786,6 +786,23 @@ static void qed_iscsi_release_connection(struct qed_hwfn *p_hwfn, spin_unlock_bh(&p_hwfn->p_iscsi_info->lock); } +void qed_iscsi_free_connection(struct qed_hwfn *p_hwfn, + struct qed_iscsi_conn *p_conn) +{ + qed_chain_free(p_hwfn->cdev, &p_conn->xhq); + qed_chain_free(p_hwfn->cdev, &p_conn->uhq); + qed_chain_free(p_hwfn->cdev, &p_conn->r2tq); + dma_free_coherent(&p_hwfn->cdev->pdev->dev, + sizeof(struct tcp_upload_params), + p_conn->tcp_upload_params_virt_addr, + p_conn->tcp_upload_params_phys_addr); + dma_free_coherent(&p_hwfn->cdev->pdev->dev, + sizeof(struct scsi_terminate_extra_params), + p_conn->queue_cnts_virt_addr, + p_conn->queue_cnts_phys_addr); + kfree(p_conn); +} + struct qed_iscsi_info *qed_iscsi_alloc(struct qed_hwfn *p_hwfn) { struct qed_iscsi_info *p_iscsi_info; @@ -807,6 +824,17 @@ void qed_iscsi_setup(struct qed_hwfn *p_hwfn, void qed_iscsi_free(struct qed_hwfn *p_hwfn, struct qed_iscsi_info *p_iscsi_info) { + struct qed_iscsi_conn *p_conn = NULL; + + while (!list_empty(&p_hwfn->p_iscsi_info->free_list)) { + p_conn = list_first_entry(&p_hwfn->p_iscsi_info->free_list, + struct qed_iscsi_conn, list_entry); + if (p_conn) { + list_del(&p_conn->list_entry); + qed_iscsi_free_connection(p_hwfn, p_conn); + } + } + kfree(p_iscsi_info); } From 1df2adedcce17ad4a39fba74f0e2b611f797fe10 Mon Sep 17 00:00:00 2001 From: Ram Amrani Date: Tue, 14 Mar 2017 15:26:02 +0200 Subject: [PATCH 5/7] qed: Fix interrupt flags on Rx LL2 Before iterating over the the LL2 Rx ring, the ring's spinlock is taken via spin_lock_irqsave(). The actual processing of the packet [including handling by the protocol driver] is done without said lock, so qed releases the spinlock and re-claims it afterwards. Problem is that the final spin_lock_irqrestore() at the end of the iteration uses the original flags saved from the initial irqsave() instead of the flags from the most recent irqsave(). So it's possible that the interrupt status would be incorrect at the end of the processing. Fixes: 0a7fb11c23c0 ("qed: Add Light L2 support"); CC: Ram Amrani Signed-off-by: Yuval Mintz Signed-off-by: David S. Miller --- drivers/net/ethernet/qlogic/qed/qed_ll2.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/drivers/net/ethernet/qlogic/qed/qed_ll2.c b/drivers/net/ethernet/qlogic/qed/qed_ll2.c index 29ae5ec2ac2db6..0d3cef409c96d0 100644 --- a/drivers/net/ethernet/qlogic/qed/qed_ll2.c +++ b/drivers/net/ethernet/qlogic/qed/qed_ll2.c @@ -476,7 +476,7 @@ qed_ll2_rxq_completion_gsi(struct qed_hwfn *p_hwfn, static int qed_ll2_rxq_completion_reg(struct qed_hwfn *p_hwfn, struct qed_ll2_info *p_ll2_conn, union core_rx_cqe_union *p_cqe, - unsigned long lock_flags, + unsigned long *p_lock_flags, bool b_last_cqe) { struct qed_ll2_rx_queue *p_rx = &p_ll2_conn->rx_queue; @@ -497,10 +497,10 @@ static int qed_ll2_rxq_completion_reg(struct qed_hwfn *p_hwfn, "Mismatch between active_descq and the LL2 Rx chain\n"); list_add_tail(&p_pkt->list_entry, &p_rx->free_descq); - spin_unlock_irqrestore(&p_rx->lock, lock_flags); + spin_unlock_irqrestore(&p_rx->lock, *p_lock_flags); qed_ll2b_complete_rx_packet(p_hwfn, p_ll2_conn->my_id, p_pkt, &p_cqe->rx_cqe_fp, b_last_cqe); - spin_lock_irqsave(&p_rx->lock, lock_flags); + spin_lock_irqsave(&p_rx->lock, *p_lock_flags); return 0; } @@ -540,7 +540,8 @@ static int qed_ll2_rxq_completion(struct qed_hwfn *p_hwfn, void *cookie) break; case CORE_RX_CQE_TYPE_REGULAR: rc = qed_ll2_rxq_completion_reg(p_hwfn, p_ll2_conn, - cqe, flags, b_last_cqe); + cqe, &flags, + b_last_cqe); break; default: rc = -EIO; From db31d330e8b0ad8926725eb2af6f6422c07ca7ab Mon Sep 17 00:00:00 2001 From: "Mintz, Yuval" Date: Tue, 14 Mar 2017 15:26:03 +0200 Subject: [PATCH 6/7] qed: Correct out-of-bound access in OOO history Need to set the number of entries in database, otherwise the logic would quickly surpass the array. Fixes: 1d6cff4fca43 ("qed: Add iSCSI out of order packet handling") Signed-off-by: Yuval Mintz Signed-off-by: David S. Miller --- drivers/net/ethernet/qlogic/qed/qed_ooo.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/net/ethernet/qlogic/qed/qed_ooo.c b/drivers/net/ethernet/qlogic/qed/qed_ooo.c index 7d731c6cb8923d..378afce58b3f0a 100644 --- a/drivers/net/ethernet/qlogic/qed/qed_ooo.c +++ b/drivers/net/ethernet/qlogic/qed/qed_ooo.c @@ -159,6 +159,8 @@ struct qed_ooo_info *qed_ooo_alloc(struct qed_hwfn *p_hwfn) if (!p_ooo_info->ooo_history.p_cqes) goto no_history_mem; + p_ooo_info->ooo_history.num_of_cqes = QED_MAX_NUM_OOO_HISTORY_ENTRIES; + return p_ooo_info; no_history_mem: From 6b116b1d6a521a1907b3c18cb7a8592a655f660c Mon Sep 17 00:00:00 2001 From: "Mintz, Yuval" Date: Tue, 14 Mar 2017 15:26:04 +0200 Subject: [PATCH 7/7] qed: Enable iSCSI Out-of-Order Missing in the initial submission, qed fails to propagate qedi's request to enable OOO to firmware. Fixes: fc831825f99e ("qed: Add support for hardware offloaded iSCSI") Signed-off-by: Yuval Mintz Signed-off-by: David S. Miller --- drivers/net/ethernet/qlogic/qed/qed_iscsi.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/net/ethernet/qlogic/qed/qed_iscsi.c b/drivers/net/ethernet/qlogic/qed/qed_iscsi.c index 7e73b05eeab862..098766f7fe88a6 100644 --- a/drivers/net/ethernet/qlogic/qed/qed_iscsi.c +++ b/drivers/net/ethernet/qlogic/qed/qed_iscsi.c @@ -190,6 +190,9 @@ qed_sp_iscsi_func_start(struct qed_hwfn *p_hwfn, p_init->num_sq_pages_in_ring = p_params->num_sq_pages_in_ring; p_init->num_r2tq_pages_in_ring = p_params->num_r2tq_pages_in_ring; p_init->num_uhq_pages_in_ring = p_params->num_uhq_pages_in_ring; + p_init->ooo_enable = p_params->ooo_enable; + p_init->ll2_rx_queue_id = p_hwfn->hw_info.resc_start[QED_LL2_QUEUE] + + p_params->ll2_ooo_queue_id; p_init->func_params.log_page_size = p_params->log_page_size; val = p_params->num_tasks; p_init->func_params.num_tasks = cpu_to_le16(val);