Merge tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma

Pull rdma fixes from Jason Gunthorpe:
 "This includes a fix for a significant security miss in checking the
  RDMA_NLDEV_CMD_SYS_SET operation.

  Summary:

   - UAF in SRP

   - Error unwind failure in siw connection management

   - Missing error checks

   - NULL/ERR_PTR confusion in erdma

   - Possible string truncation in CMA configfs and mlx4

   - Data ordering issue in bnxt_re

   - Missing stats decrement on object destroy in bnxt_re

   - Mlx5 bugs in this merge window:
      * Incorrect access_flag in the new mkey cache
      * Missing unlock on error in flow steering
      * lockdep possible deadlock on new mkey cache destruction (Plus a
        fix for this too)

   - Don't leak kernel stack memory to userspace in the CM

   - Missing permission validation for RDMA_NLDEV_CMD_SYS_SET"

* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma:
  RDMA/core: Require admin capabilities to set system parameters
  RDMA/mlx5: Remove not-used cache disable flag
  RDMA/cma: Initialize ib_sa_multicast structure to 0 when join
  RDMA/mlx5: Fix mkey cache possible deadlock on cleanup
  RDMA/mlx5: Fix NULL string error
  RDMA/mlx5: Fix mutex unlocking on error flow for steering anchor creation
  RDMA/mlx5: Fix assigning access flags to cache mkeys
  IB/mlx4: Fix the size of a buffer in add_port_entries()
  RDMA/bnxt_re: Decrement resource stats correctly
  RDMA/bnxt_re: Fix the handling of control path response data
  RDMA/cma: Fix truncation compilation warning in make_cma_ports
  RDMA/erdma: Fix NULL pointer access in regmr_cmd
  RDMA/erdma: Fix error code in erdma_create_scatter_mtt()
  RDMA/uverbs: Fix typo of sizeof argument
  RDMA/cxgb4: Check skb value for failure to allocate
  RDMA/siw: Fix connection failure handling
  RDMA/srp: Do not call scsi_done() from srp_abort()

drivers/infiniband/core/cma.c

@@ -4968,7 +4968,7 @@ static int cma_iboe_join_multicast(struct rdma_id_private *id_priv,
 	int err = 0;
 	struct sockaddr *addr = (struct sockaddr *)&mc->addr;
 	struct net_device *ndev = NULL;
-	struct ib_sa_multicast ib;
+	struct ib_sa_multicast ib = {};
 	enum ib_gid_type gid_type;
 	bool send_only;
 

drivers/infiniband/core/cma_configfs.c

@@ -217,7 +217,7 @@ static int make_cma_ports(struct cma_dev_group *cma_dev_group,
 		return -ENOMEM;
 
 	for (i = 0; i < ports_num; i++) {
-		char port_str[10];
+		char port_str[11];
 
 		ports[i].port_num = i + 1;
 		snprintf(port_str, sizeof(port_str), "%u", i + 1);

drivers/infiniband/core/nldev.c

@@ -2529,6 +2529,7 @@ static const struct rdma_nl_cbs nldev_cb_table[RDMA_NLDEV_NUM_OPS] = {
 	},
 	[RDMA_NLDEV_CMD_SYS_SET] = {
 		.doit = nldev_set_sys_set_doit,
+		.flags = RDMA_NL_ADMIN_PERM,
 	},
 	[RDMA_NLDEV_CMD_STAT_SET] = {
 		.doit = nldev_stat_set_doit,

drivers/infiniband/core/uverbs_main.c

@@ -546,7 +546,7 @@ static ssize_t verify_hdr(struct ib_uverbs_cmd_hdr *hdr,
 	if (hdr->in_words * 4 != count)
 		return -EINVAL;
 
-	if (count < method_elm->req_size + sizeof(hdr)) {
+	if (count < method_elm->req_size + sizeof(*hdr)) {
 		/*
 		 * rdma-core v18 and v19 have a bug where they send DESTROY_CQ
 		 * with a 16 byte write instead of 24. Old kernels didn't

drivers/infiniband/hw/bnxt_re/ib_verbs.c

@@ -910,6 +910,10 @@ int bnxt_re_destroy_qp(struct ib_qp *ib_qp, struct ib_udata *udata)
 	list_del(&qp->list);
 	mutex_unlock(&rdev->qp_lock);
 	atomic_dec(&rdev->stats.res.qp_count);
+	if (qp->qplib_qp.type == CMDQ_CREATE_QP_TYPE_RC)
+		atomic_dec(&rdev->stats.res.rc_qp_count);
+	else if (qp->qplib_qp.type == CMDQ_CREATE_QP_TYPE_UD)
+		atomic_dec(&rdev->stats.res.ud_qp_count);
 
 	ib_umem_release(qp->rumem);
 	ib_umem_release(qp->sumem);

drivers/infiniband/hw/bnxt_re/qplib_rcfw.c

@@ -665,7 +665,6 @@ static int bnxt_qplib_process_qp_event(struct bnxt_qplib_rcfw *rcfw,
 		blocked = cookie & RCFW_CMD_IS_BLOCKING;
 		cookie &= RCFW_MAX_COOKIE_VALUE;
 		crsqe = &rcfw->crsqe_tbl[cookie];
-		crsqe->is_in_used = false;
 
 		if (WARN_ONCE(test_bit(FIRMWARE_STALL_DETECTED,
 				       &rcfw->cmdq.flags),
@@ -681,8 +680,14 @@ static int bnxt_qplib_process_qp_event(struct bnxt_qplib_rcfw *rcfw,
 			atomic_dec(&rcfw->timeout_send);
 
 		if (crsqe->is_waiter_alive) {
-			if (crsqe->resp)
+			if (crsqe->resp) {
 				memcpy(crsqe->resp, qp_event, sizeof(*qp_event));
+				/* Insert write memory barrier to ensure that
+				 * response data is copied before clearing the
+				 * flags
+				 */
+				smp_wmb();
+			}
 			if (!blocked)
 				wait_cmds++;
 		}
@@ -694,6 +699,8 @@ static int bnxt_qplib_process_qp_event(struct bnxt_qplib_rcfw *rcfw,
 		if (!is_waiter_alive)
 			crsqe->resp = NULL;
 
+		crsqe->is_in_used = false;
+
 		hwq->cons += req_size;
 
 		/* This is a case to handle below scenario -

drivers/infiniband/hw/cxgb4/cm.c

@@ -1965,6 +1965,9 @@ static int send_fw_act_open_req(struct c4iw_ep *ep, unsigned int atid)
 	int win;
 
 	skb = get_skb(NULL, sizeof(*req), GFP_KERNEL);
+	if (!skb)
+		return -ENOMEM;
+
 	req = __skb_put_zero(skb, sizeof(*req));
 	req->op_compl = htonl(WR_OP_V(FW_OFLD_CONNECTION_WR));
 	req->len16_pkd = htonl(FW_WR_LEN16_V(DIV_ROUND_UP(sizeof(*req), 16)));

drivers/infiniband/hw/erdma/erdma_verbs.c

@@ -133,8 +133,8 @@ static int create_qp_cmd(struct erdma_ucontext *uctx, struct erdma_qp *qp)
 static int regmr_cmd(struct erdma_dev *dev, struct erdma_mr *mr)
 {
 	struct erdma_pd *pd = to_epd(mr->ibmr.pd);
+	u32 mtt_level = ERDMA_MR_MTT_0LEVEL;
 	struct erdma_cmdq_reg_mr_req req;
-	u32 mtt_level;
 
 	erdma_cmdq_build_reqhdr(&req.hdr, CMDQ_SUBMOD_RDMA, CMDQ_OPCODE_REG_MR);
 
@@ -147,10 +147,9 @@ static int regmr_cmd(struct erdma_dev *dev, struct erdma_mr *mr)
 			req.phy_addr[0] = sg_dma_address(mr->mem.mtt->sglist);
 			mtt_level = mr->mem.mtt->level;
 		}
-	} else {
+	} else if (mr->type != ERDMA_MR_TYPE_DMA) {
 		memcpy(req.phy_addr, mr->mem.mtt->buf,
 		       MTT_SIZE(mr->mem.page_cnt));
-		mtt_level = ERDMA_MR_MTT_0LEVEL;
 	}
 
 	req.cfg0 = FIELD_PREP(ERDMA_CMD_MR_VALID_MASK, mr->valid) |
@@ -655,7 +654,7 @@ static struct erdma_mtt *erdma_create_scatter_mtt(struct erdma_dev *dev,
 
 	mtt = kzalloc(sizeof(*mtt), GFP_KERNEL);
 	if (!mtt)
-		return NULL;
+		return ERR_PTR(-ENOMEM);
 
 	mtt->size = ALIGN(size, PAGE_SIZE);
 	mtt->buf = vzalloc(mtt->size);

drivers/infiniband/hw/mlx4/sysfs.c

@@ -223,7 +223,7 @@ void del_sysfs_port_mcg_attr(struct mlx4_ib_dev *device, int port_num,
 static int add_port_entries(struct mlx4_ib_dev *device, int port_num)
 {
 	int i;
-	char buff[11];
+	char buff[12];
 	struct mlx4_ib_iov_port *port = NULL;
 	int ret = 0 ;
 	struct ib_port_attr attr;

drivers/infiniband/hw/mlx5/fs.c

@@ -2470,8 +2470,8 @@ static int UVERBS_HANDLER(MLX5_IB_METHOD_STEERING_ANCHOR_CREATE)(
 	mlx5_steering_anchor_destroy_res(ft_prio);
 put_flow_table:
 	put_flow_table(dev, ft_prio, true);
-	mutex_unlock(&dev->flow_db->lock);
 free_obj:
+	mutex_unlock(&dev->flow_db->lock);
 	kfree(obj);
 
 	return err;

drivers/infiniband/hw/mlx5/main.c

@@ -2084,7 +2084,7 @@ static inline char *mmap_cmd2str(enum mlx5_ib_mmap_cmd cmd)
 	case MLX5_IB_MMAP_DEVICE_MEM:
 		return "Device Memory";
 	default:
-		return NULL;
+		return "Unknown";
 	}
 }
 

drivers/infiniband/hw/mlx5/mr.c

@@ -301,7 +301,8 @@ static int get_mkc_octo_size(unsigned int access_mode, unsigned int ndescs)
 
 static void set_cache_mkc(struct mlx5_cache_ent *ent, void *mkc)
 {
-	set_mkc_access_pd_addr_fields(mkc, 0, 0, ent->dev->umrc.pd);
+	set_mkc_access_pd_addr_fields(mkc, ent->rb_key.access_flags, 0,
+				      ent->dev->umrc.pd);
 	MLX5_SET(mkc, mkc, free, 1);
 	MLX5_SET(mkc, mkc, umr_en, 1);
 	MLX5_SET(mkc, mkc, access_mode_1_0, ent->rb_key.access_mode & 0x3);
@@ -1024,19 +1025,26 @@ void mlx5_mkey_cache_cleanup(struct mlx5_ib_dev *dev)
 	if (!dev->cache.wq)
 		return;
 
-	cancel_delayed_work_sync(&dev->cache.remove_ent_dwork);
 	mutex_lock(&dev->cache.rb_lock);
 	for (node = rb_first(root); node; node = rb_next(node)) {
 		ent = rb_entry(node, struct mlx5_cache_ent, node);
 		xa_lock_irq(&ent->mkeys);
 		ent->disabled = true;
 		xa_unlock_irq(&ent->mkeys);
-		cancel_delayed_work_sync(&ent->dwork);
 	}
+	mutex_unlock(&dev->cache.rb_lock);
+
+	/*
+	 * After all entries are disabled and will not reschedule on WQ,
+	 * flush it and all async commands.
+	 */
+	flush_workqueue(dev->cache.wq);
 
 	mlx5_mkey_cache_debugfs_cleanup(dev);
 	mlx5_cmd_cleanup_async_ctx(&dev->async_ctx);
 
+	/* At this point all entries are disabled and have no concurrent work. */
+	mutex_lock(&dev->cache.rb_lock);
 	node = rb_first(root);
 	while (node) {
 		ent = rb_entry(node, struct mlx5_cache_ent, node);

drivers/infiniband/sw/siw/siw_cm.c

@@ -976,6 +976,7 @@ static void siw_accept_newconn(struct siw_cep *cep)
 			siw_cep_put(cep);
 			new_cep->listen_cep = NULL;
 			if (rv) {
+				siw_cancel_mpatimer(new_cep);
 				siw_cep_set_free(new_cep);
 				goto error;
 			}
@@ -1100,11 +1101,14 @@ static void siw_cm_work_handler(struct work_struct *w)
 				/*
 				 * Socket close before MPA request received.
 				 */
-				siw_dbg_cep(cep, "no mpareq: drop listener\n");
+				if (cep->listen_cep) {
+					siw_dbg_cep(cep,
+						"no mpareq: drop listener\n");
 					siw_cep_put(cep->listen_cep);
 					cep->listen_cep = NULL;
 				}
 			}
+		}
 		release_cep = 1;
 		break;
 
@@ -1227,7 +1231,11 @@ static void siw_cm_llp_data_ready(struct sock *sk)
 	if (!cep)
 		goto out;
 
-	siw_dbg_cep(cep, "state: %d\n", cep->state);
+	siw_dbg_cep(cep, "cep state: %d, socket state %d\n",
+		    cep->state, sk->sk_state);
+
+	if (sk->sk_state != TCP_ESTABLISHED)
+		goto out;
 
 	switch (cep->state) {
 	case SIW_EPSTATE_RDMA_MODE:

drivers/infiniband/ulp/srp/ib_srp.c

@@ -2784,7 +2784,6 @@ static int srp_abort(struct scsi_cmnd *scmnd)
 	u32 tag;
 	u16 ch_idx;
 	struct srp_rdma_ch *ch;
-	int ret;
 
 	shost_printk(KERN_ERR, target->scsi_host, "SRP abort called\n");
 
@@ -2798,19 +2797,14 @@ static int srp_abort(struct scsi_cmnd *scmnd)
 	shost_printk(KERN_ERR, target->scsi_host,
 		     "Sending SRP abort for tag %#x\n", tag);
 	if (srp_send_tsk_mgmt(ch, tag, scmnd->device->lun,
-			      SRP_TSK_ABORT_TASK, NULL) == 0)
-		ret = SUCCESS;
-	else if (target->rport->state == SRP_RPORT_LOST)
-		ret = FAST_IO_FAIL;
-	else
-		ret = FAILED;
-	if (ret == SUCCESS) {
+			      SRP_TSK_ABORT_TASK, NULL) == 0) {
 		srp_free_req(ch, req, scmnd, 0);
-		scmnd->result = DID_ABORT << 16;
-		scsi_done(scmnd);
+		return SUCCESS;
 	}
+	if (target->rport->state == SRP_RPORT_LOST)
+		return FAST_IO_FAIL;
 
-	return ret;
+	return FAILED;
 }
 
 static int srp_reset_device(struct scsi_cmnd *scmnd)