Skip to contentSkip to conversation

net/mlx5: fix ROCE LAG bond device probing

PATCH Applied Viacheslav Ovsiienko 2 participants 2 messages Jul 21, 2021 – Jul 22, 2021 Network drivers

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 net/mlx5: fix ROCE LAG bond device probing

The ROCE LAG bond device requires neither E-Switch nor SR-IOV configurations. It means the ROCE LAG bond device might be presented as a single port Infiniband device.

The mlx5 PMD wrongly recognized standalone ROCE LAG bond device as E-Switch configuration, this triggered the calls of E-Switch ports related API and the latter failed (over the new OFED kernel driver, starting since 5.4.1), causing the overall device probe failure.

If there is a single port Infiniband bond device found the E-Switch related flags must be cleared indicating standalone configuration.

Also, it is not true anymore the bond device can exist over E-Switch configurations only (as it was claimed for VF LAG bond devices). The related checks are not relevant anymore and removed.

Fixes 790164ce1d2d ("net/mlx5: check kernel support for VF LAG bonding")
---
 drivers/net/mlx5/linux/mlx5_os.c | 24 +++++++++---------------
 1 file changed, 9 insertions(+), 15 deletions(-)
drivers/net/mlx5/linux/mlx5_os.c +9 -15
index aa5210fa45..e568cc9c48 100644
@@ -2216,19 +2216,6 @@ mlx5_os_pci_probe_pf(struct rte_pci_device *pci_dev,
 			goto exit;
 		}
 	}
-#ifndef HAVE_MLX5DV_DR_DEVX_PORT
-	if (bd >= 0) {
-		/*
-		 * This may happen if there is VF LAG kernel support and
-		 * application is compiled with older rdma_core library.
-		 */
-		DRV_LOG(ERR,
-			"No kernel/verbs support for VF LAG bonding found.");
-		rte_errno = ENOTSUP;
-		ret = -rte_errno;
-		goto exit;
-	}
-#endif
 	/*
 	 * Now we can determine the maximal
 	 * amount of devices to be spawned.
@@ -2292,10 +2279,18 @@ mlx5_os_pci_probe_pf(struct rte_pci_device *pci_dev,
 						(list[ns].ifindex,
 						 &list[ns].info);
 			}
-#ifdef HAVE_MLX5DV_DR_DEVX_PORT
 			if (!ret && bd >= 0) {
 				switch (list[ns].info.name_type) {
 				case MLX5_PHYS_PORT_NAME_TYPE_UPLINK:
+					if (np == 1) {
+						/*
+						 * Force standalone bonding
+						 * device for ROCE LAG
+						 * confgiurations.
+						 */
+						list[ns].info.master = 0;
+						list[ns].info.representor = 0;
+					}
 					if (list[ns].info.port_name == bd)
 						ns++;
 					break;
@@ -2312,7 +2307,6 @@ mlx5_os_pci_probe_pf(struct rte_pci_device *pci_dev,
 				}
 				continue;
 			}
-#endif
 			if (!ret && (list[ns].info.representor ^
 				     list[ns].info.master))
 				ns++;
Signature · 1 line
2.18.1
#1 lore raw

21/07/2021 10:31, Viacheslav Ovsiienko:

Quoted 19 lines

The ROCE LAG bond device requires neither E-Switch nor SR-IOV configurations. It means the ROCE LAG bond device might be presented as a single port Infiniband device.

The mlx5 PMD wrongly recognized standalone ROCE LAG bond device as E-Switch configuration, this triggered the calls of E-Switch ports related API and the latter failed (over the new OFED kernel driver, starting since 5.4.1), causing the overall device probe failure.

If there is a single port Infiniband bond device found the E-Switch related flags must be cleared indicating standalone configuration.

Also, it is not true anymore the bond device can exist over E-Switch configurations only (as it was claimed for VF LAG bond devices). The related checks are not relevant anymore and removed.

Quoted 5 lines
Fixes 790164ce1d2d ("net/mlx5: check kernel support for VF LAG bonding")

Applied, thanks