Skip to contentSkip to conversation

nvme-multipath: find NUMA path only for online numa-node

PATCH New Sasha Levin 1 participant 2 messages Jun 5, 2024

Subsystems (2)
lore
Share & tools
lore

Sasha Levin

author maintainer
#0 lore raw

PATCH 1/2 nvme-multipath: find NUMA path only for online numa-node

From: Nilay Shroff <[email protected]>

[ Upstream commit d3a043733f25d743f3aa617c7f82dbcb5ee2211a ]

In current native multipath design when a shared namespace is created, we loop through each possible numa-node, calculate the NUMA distance of that node from each nvme controller and then cache the optimal IO path for future reference while sending IO. The issue with this design is that we may refer to the NUMA distance table for an offline node which may not be populated at the time and so we may inadvertently end up finding and caching a non-optimal path for IO. Then latter when the corresponding numa-node becomes online and hence the NUMA distance table entry for that node is created, ideally we should re-calculate the multipath node distance for the newly added node however that doesn't happen unless we rescan/reset the controller. So essentially, we may keep using non-optimal IO path for a node which is made online after namespace is created.
This patch helps fix this issue ensuring that when a shared namespace is created, we calculate the multipath node distance for each online numa-node instead of each possible numa-node. Then latter when a node becomes online and we receive any IO on that newly added node, we would calculate the multipath node distance for newly added node but this time NUMA distance table would have been already populated for newly added node. Hence we would be able to correctly calculate the multipath node distance and choose the optimal path for the IO.

---
 drivers/nvme/host/multipath.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/nvme/host/multipath.c +1 -1
index 811f7b96b5517..c993548403f5c 100644
@@ -436,7 +436,7 @@ static void nvme_mpath_set_live(struct nvme_ns *ns)
 		int node, srcu_idx;
 
 		srcu_idx = srcu_read_lock(&head->srcu);
-		for_each_node(node)
+		for_each_online_node(node)
 			__nvme_find_path(head, node);
 		srcu_read_unlock(&head->srcu, srcu_idx);
 	}
Signature · 1 line
2.43.0
author maintainer
#1 lore raw

PATCH 2/2 afs: Don't cross .backup mountpoint from backup volume

From: Marc Dionne <[email protected]>

[ Upstream commit 29be9100aca2915fab54b5693309bc42956542e5 ]

Don't cross a mountpoint that explicitly specifies a backup volume (target is <vol>.backup) when starting from a backup volume.

It it not uncommon to mount a volume's backup directly in the volume itself. This can cause tools that are not paying attention to get into a loop mounting the volume onto itself as they attempt to traverse the tree, leading to a variety of problems.

This doesn't prevent the general case of loops in a sequence of mountpoints, but addresses a common special case in the same way as other afs clients.

---
 fs/afs/mntpt.c | 5 +++++
 1 file changed, 5 insertions(+)
fs/afs/mntpt.c +5 -0
index f105f061e89bb..eeb2b4b11e323 100644
@@ -146,6 +146,11 @@ static int afs_mntpt_set_params(struct fs_context *fc, struct dentry *mntpt)
 		put_page(page);
 		if (ret < 0)
 			return ret;
+
+		/* Don't cross a backup volume mountpoint from a backup volume */
+		if (src_as->volume && src_as->volume->type == AFSVL_BACKVOL &&
+		    ctx->type == AFSVL_BACKVOL)
+			return -ENODEV;
 	}
 
 	return 0;
Signature · 1 line
2.43.0