The routine is always called towards the end of lookup.

According to bpftrace on my boxen and boxen of people I asked, the depth
count is almost always 0, thus the call can be avoided in the common case.

one-liner:
bpftrace -e 'kprobe:legitimize_links { @[((struct nameidata *)arg0)->depth] = count(); }'

sample results from few minutes of tracing:
@[1]: 59
@[0]: 147236

@[2]: 1
@[1]: 12087
@[0]: 5926235

And of course the venerable kernel build:
@[1]: 3563
@[0]: 6625425

Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>
---
 fs/namei.c | 19 ++++++++++++++-----
 1 file changed, 14 insertions(+), 5 deletions(-)

diff --git a/fs/namei.c b/fs/namei.c
index 2a112b2c0951..d89937c2c0b2 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -826,7 +826,7 @@ static inline bool legitimize_path(struct nameidata *nd,
 	return __legitimize_path(path, seq, nd->m_seq);
 }
 
-static bool legitimize_links(struct nameidata *nd)
+static noinline bool legitimize_links(struct nameidata *nd)
 {
 	int i;
 	if (unlikely(nd->flags & LOOKUP_CACHED)) {
@@ -845,6 +845,11 @@ static bool legitimize_links(struct nameidata *nd)
 	return true;
 }
 
+static __always_inline bool need_legitimize_links(struct nameidata *nd)
+{
+	return nd->depth > 0;
+}
+
 static bool legitimize_root(struct nameidata *nd)
 {
 	/* Nothing to do if nd->root is zero or is managed by the VFS user. */
@@ -882,8 +887,10 @@ static bool try_to_unlazy(struct nameidata *nd)
 
 	BUG_ON(!(nd->flags & LOOKUP_RCU));
 
-	if (unlikely(!legitimize_links(nd)))
-		goto out1;
+	if (unlikely(need_legitimize_links(nd))) {
+		if (unlikely(!legitimize_links(nd)))
+			goto out1;
+	}
 	if (unlikely(!legitimize_path(nd, &nd->path, nd->seq)))
 		goto out;
 	if (unlikely(!legitimize_root(nd)))
@@ -917,8 +924,10 @@ static bool try_to_unlazy_next(struct nameidata *nd, struct dentry *dentry)
 	int res;
 	BUG_ON(!(nd->flags & LOOKUP_RCU));
 
-	if (unlikely(!legitimize_links(nd)))
-		goto out2;
+	if (unlikely(need_legitimize_links(nd))) {
+		if (unlikely(!legitimize_links(nd)))
+			goto out2;
+	}
 	res = __legitimize_mnt(nd->path.mnt, nd->m_seq);
 	if (unlikely(res)) {
 		if (res > 0)
-- 
2.48.1

On Sun, Nov 09, 2025 at 07:54:09PM +0100, Mateusz Guzik wrote:

> @@ -882,8 +887,10 @@ static bool try_to_unlazy(struct nameidata *nd)
>  
>  	BUG_ON(!(nd->flags & LOOKUP_RCU));
>  
> -	if (unlikely(!legitimize_links(nd)))
> -		goto out1;
> +	if (unlikely(need_legitimize_links(nd))) {
> +		if (unlikely(!legitimize_links(nd)))
> +			goto out1;
> +	}
>  	if (unlikely(!legitimize_path(nd, &nd->path, nd->seq)))
>  		goto out;
>  	if (unlikely(!legitimize_root(nd)))
> @@ -917,8 +924,10 @@ static bool try_to_unlazy_next(struct nameidata *nd, struct dentry *dentry)
>  	int res;
>  	BUG_ON(!(nd->flags & LOOKUP_RCU));
>  
> -	if (unlikely(!legitimize_links(nd)))
> -		goto out2;
> +	if (unlikely(need_legitimize_links(nd))) {
> +		if (unlikely(!legitimize_links(nd)))
> +			goto out2;
> +	}
>  	res = __legitimize_mnt(nd->path.mnt, nd->m_seq);
>  	if (unlikely(res)) {
>  		if (res > 0)

Seeing that odds of extra callers showing up are pretty much nil,
I think your need_legitimize_links() is only obfuscating things.
Let's just make it
	if (unlikely(nd->depth) && !legitimize_links(nd))
		goto ...
and be done with that.