[PATCH] Fix get_node_by_path string equality check

[PATCH] Fix get_node_by_path string equality check

[Tim Montague]

When determining if to recurse into a node, get_node_by_path does not
check if the length of each node name is equal.  If searching for
/foo/baz, this can result in recursing into /food because
strneq("foo", "food", 3) is true.

Signed-off-by: Tim Montague <tmontague@ghs.com>
---
 livetree.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/livetree.c b/livetree.c
index 3673de0..aecd278 100644
--- a/livetree.c
+++ b/livetree.c
@@ -478,7 +478,8 @@ struct node *get_node_by_path(struct node *tree, const char *path)
 	p = strchr(path, '/');
 
 	for_each_child(tree, child) {
-		if (p && strneq(path, child->name, p-path))
+		if (p && (strlen(child->name) == p-path) &&
+				strneq(path, child->name, p-path))
 			return get_node_by_path(child, p+1);
 		else if (!p && streq(path, child->name))
 			return child;
-- 
2.7.4

Re: [PATCH] Fix get_node_by_path string equality check

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 1339 bytes --]

On Tue, Apr 18, 2017 at 12:08:18AM +0000, Tim Montague wrote:
> When determining if to recurse into a node, get_node_by_path does not
> check if the length of each node name is equal.  If searching for
> /foo/baz, this can result in recursing into /food because
> strneq("foo", "food", 3) is true.
> 
> Signed-off-by: Tim Montague <tmontague-5j0BnJlmnLM@public.gmane.org>

Looks like I made the classic strncmp() blunder, how embarrassing.

Fix looks good, but I could trouble you to add a testcase exercising
this?


> ---
>  livetree.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/livetree.c b/livetree.c
> index 3673de0..aecd278 100644
> --- a/livetree.c
> +++ b/livetree.c
> @@ -478,7 +478,8 @@ struct node *get_node_by_path(struct node *tree, const char *path)
>  	p = strchr(path, '/');
>  
>  	for_each_child(tree, child) {
> -		if (p && strneq(path, child->name, p-path))
> +		if (p && (strlen(child->name) == p-path) &&
> +				strneq(path, child->name, p-path))
>  			return get_node_by_path(child, p+1);
>  		else if (!p && streq(path, child->name))
>  			return child;

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

RE: [PATCH] Fix get_node_by_path string equality check

[Tim Montague]

When determining if to recurse into a node, get_node_by_path does not
check if the length of each node name is equal.  If searching for
/foo/baz, this can result in recursing into /foobar because
strneq("foo", "foobar", 3) is true.

This can result in a reference to /foo/baz to be incorrectly set to
/foobar/baz.  A test for this was added.

Signed-off-by: Tim Montague <tmontague@ghs.com>
---
 livetree.c                |  3 ++-
 tests/path-references.c   | 12 +++++++++++-
 tests/path-references.dts | 13 +++++++++++++
 3 files changed, 26 insertions(+), 2 deletions(-)

diff --git a/livetree.c b/livetree.c
index 3673de0..aecd278 100644
--- a/livetree.c
+++ b/livetree.c
@@ -478,7 +478,8 @@ struct node *get_node_by_path(struct node *tree, const char *path)
 	p = strchr(path, '/');
 
 	for_each_child(tree, child) {
-		if (p && strneq(path, child->name, p-path))
+		if (p && (strlen(child->name) == p-path) &&
+				strneq(path, child->name, p-path))
 			return get_node_by_path(child, p+1);
 		else if (!p && streq(path, child->name))
 			return child;
diff --git a/tests/path-references.c b/tests/path-references.c
index c8d25fb..5e332e8 100644
--- a/tests/path-references.c
+++ b/tests/path-references.c
@@ -66,7 +66,7 @@ int main(int argc, char *argv[])
 	void *fdt;
 	const char *p;
 	int len, multilen;
-	int n1, n2;
+	int n1, n2, n3, n4;
 
 	test_init(argc, argv);
 	fdt = load_blob_arg(argc, argv);
@@ -92,6 +92,16 @@ int main(int argc, char *argv[])
 	if ((!streq(p, "/node1") || !streq(p + strlen("/node1") + 1, "/node2")))
 		FAIL("multiref has wrong value");
 
+	/* Check reference to nested nodes with common prefix */
+	n3 = fdt_path_offset(fdt, "/foo/baz");
+	if (n3 < 0)
+		FAIL("fdt_path_offset(/foo/baz): %s", fdt_strerror(n3));
+	n4 = fdt_path_offset(fdt, "/foobar/baz");
+	if (n4 < 0)
+		FAIL("fdt_path_offset(/foobar/baz): %s", fdt_strerror(n4));
+	check_ref(fdt, n3, "/foobar/baz");
+	check_ref(fdt, n4, "/foo/baz");
+
 	check_rref(fdt);
 
 	PASS();
diff --git a/tests/path-references.dts b/tests/path-references.dts
index b00fd79..3808aa6 100644
--- a/tests/path-references.dts
+++ b/tests/path-references.dts
@@ -12,4 +12,17 @@
 		ref = &{/node1}; /* reference after target */
 		lref = &n1;
 	};
+	/* Check references to nested nodes with common prefix */
+	foobar {
+		n3: baz {
+		    ref = &{/foo/baz};
+		    lref = &n4;
+		};
+	};
+	foo {
+		n4: baz {
+		    ref = &{/foobar/baz};
+		    lref = &n3;
+		};
+	};
 };
-- 
2.7.4

Re: [PATCH] Fix get_node_by_path string equality check

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 4340 bytes --]

On Tue, Apr 18, 2017 at 03:18:55AM +0000, Tim Montague wrote:
> When determining if to recurse into a node, get_node_by_path does not
> check if the length of each node name is equal.  If searching for
> /foo/baz, this can result in recursing into /foobar because
> strneq("foo", "foobar", 3) is true.
> 
> This can result in a reference to /foo/baz to be incorrectly set to
> /foobar/baz.  A test for this was added.
> 
> Signed-off-by: Tim Montague <tmontague-5j0BnJlmnLM@public.gmane.org>

Ah.. it looks like your mailer has mangled the patch, I'm afraid:

$ git am -s ~/foobox/
Applying: Fix get_node_by_path string equality check
.git/rebase-apply/patch:16: trailing whitespace.
		if (p && (strlen(child->name) == p-path) &&
.git/rebase-apply/patch:17: trailing whitespace.
				strneq(path, child->name, p-path))
.git/rebase-apply/patch:30: trailing whitespace.
	int n1, n2, n3, n4;
.git/rebase-apply/patch:38: trailing whitespace.
	/* Check reference to nested nodes with common prefix */
.git/rebase-apply/patch:39: trailing whitespace.
	n3 = fdt_path_offset(fdt, "/foo/baz");
error: patch failed: livetree.c:478
error: livetree.c: patch does not apply
error: patch failed: tests/path-references.c:66
error: tests/path-references.c: patch does not apply
error: patch failed: tests/path-references.dts:12
error: tests/path-references.dts: patch does not apply
Patch failed at 0001 Fix get_node_by_path string equality check
The copy of the patch that failed is found in: .git/rebase-apply/patch
When you have resolved this problem, run "git am --continue".
If you prefer to skip this patch, run "git am --skip" instead.
To restore the original branch and stop patching, run "git am --abort".

Can you check your mailer settings.  If it's not clear how to stop it
doing that, it's probably safe to attach the patch instead.

> ---
>  livetree.c                |  3 ++-
>  tests/path-references.c   | 12 +++++++++++-
>  tests/path-references.dts | 13 +++++++++++++
>  3 files changed, 26 insertions(+), 2 deletions(-)
> 
> diff --git a/livetree.c b/livetree.c
> index 3673de0..aecd278 100644
> --- a/livetree.c
> +++ b/livetree.c
> @@ -478,7 +478,8 @@ struct node *get_node_by_path(struct node *tree, const char *path)
>  	p = strchr(path, '/');
>  
>  	for_each_child(tree, child) {
> -		if (p && strneq(path, child->name, p-path))
> +		if (p && (strlen(child->name) == p-path) &&
> +				strneq(path, child->name, p-path))
>  			return get_node_by_path(child, p+1);
>  		else if (!p && streq(path, child->name))
>  			return child;
> diff --git a/tests/path-references.c b/tests/path-references.c
> index c8d25fb..5e332e8 100644
> --- a/tests/path-references.c
> +++ b/tests/path-references.c
> @@ -66,7 +66,7 @@ int main(int argc, char *argv[])
>  	void *fdt;
>  	const char *p;
>  	int len, multilen;
> -	int n1, n2;
> +	int n1, n2, n3, n4;
>  
>  	test_init(argc, argv);
>  	fdt = load_blob_arg(argc, argv);
> @@ -92,6 +92,16 @@ int main(int argc, char *argv[])
>  	if ((!streq(p, "/node1") || !streq(p + strlen("/node1") + 1, "/node2")))
>  		FAIL("multiref has wrong value");
>  
> +	/* Check reference to nested nodes with common prefix */
> +	n3 = fdt_path_offset(fdt, "/foo/baz");
> +	if (n3 < 0)
> +		FAIL("fdt_path_offset(/foo/baz): %s", fdt_strerror(n3));
> +	n4 = fdt_path_offset(fdt, "/foobar/baz");
> +	if (n4 < 0)
> +		FAIL("fdt_path_offset(/foobar/baz): %s", fdt_strerror(n4));
> +	check_ref(fdt, n3, "/foobar/baz");
> +	check_ref(fdt, n4, "/foo/baz");
> +
>  	check_rref(fdt);
>  
>  	PASS();
> diff --git a/tests/path-references.dts b/tests/path-references.dts
> index b00fd79..3808aa6 100644
> --- a/tests/path-references.dts
> +++ b/tests/path-references.dts
> @@ -12,4 +12,17 @@
>  		ref = &{/node1}; /* reference after target */
>  		lref = &n1;
>  	};
> +	/* Check references to nested nodes with common prefix */
> +	foobar {
> +		n3: baz {
> +		    ref = &{/foo/baz};
> +		    lref = &n4;
> +		};
> +	};
> +	foo {
> +		n4: baz {
> +		    ref = &{/foobar/baz};
> +		    lref = &n3;
> +		};
> +	};
>  };

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

RE: [PATCH] Fix get_node_by_path string equality check

[Tim Montague]

[-- Attachment #1: Type: text/plain, Size: 1504 bytes --]

 > Ah.. it looks like your mailer has mangled the patch, I'm afraid:
> 
> $ git am -s ~/foobox/
> Applying: Fix get_node_by_path string equality check
> .git/rebase-apply/patch:16: trailing whitespace.
> 		if (p && (strlen(child->name) == p-path) &&
> .git/rebase-apply/patch:17: trailing whitespace.
> 				strneq(path, child->name, p-path))
> .git/rebase-apply/patch:30: trailing whitespace.
> 	int n1, n2, n3, n4;
> .git/rebase-apply/patch:38: trailing whitespace.
> 	/* Check reference to nested nodes with common prefix */
> .git/rebase-apply/patch:39: trailing whitespace.
> 	n3 = fdt_path_offset(fdt, "/foo/baz");
> error: patch failed: livetree.c:478
> error: livetree.c: patch does not apply
> error: patch failed: tests/path-references.c:66
> error: tests/path-references.c: patch does not apply
> error: patch failed: tests/path-references.dts:12
> error: tests/path-references.dts: patch does not apply
> Patch failed at 0001 Fix get_node_by_path string equality check
> The copy of the patch that failed is found in: .git/rebase-apply/patch
> When you have resolved this problem, run "git am --continue".
> If you prefer to skip this patch, run "git am --skip" instead.
> To restore the original branch and stop patching, run "git am --abort".
> 
> Can you check your mailer settings.  If it's not clear how to stop it
> doing that, it's probably safe to attach the patch instead.

Sorry about that, Outlook doesn't always behave. The patch is attached.

-Tim

[-- Attachment #2: 0001-Fix-get_node_by_path-string-equality-check.patch --]
[-- Type: application/octet-stream, Size: 2657 bytes --]

From 2bdd399513498a65913996638fdfd59c07704db8 Mon Sep 17 00:00:00 2001
From: Tim Montague <tmontague@ghs.com>
Date: Mon, 17 Apr 2017 16:51:05 -0700
Subject: [PATCH] Fix get_node_by_path string equality check

When determining if to recurse into a node, get_node_by_path does not
check if the length of each node name is equal.  If searching for
/foo/baz, this can result in recursing into /foobar because
strneq("foo", "foobar", 3) is true.

This can result in a reference to /foo/baz to be incorrectly set to
/foobar/baz.  A test for this was added.

Signed-off-by: Tim Montague <tmontague@ghs.com>
---
 livetree.c                |  3 ++-
 tests/path-references.c   | 12 +++++++++++-
 tests/path-references.dts | 13 +++++++++++++
 3 files changed, 26 insertions(+), 2 deletions(-)

diff --git a/livetree.c b/livetree.c
index 3673de0..aecd278 100644
--- a/livetree.c
+++ b/livetree.c
@@ -478,7 +478,8 @@ struct node *get_node_by_path(struct node *tree, const char *path)
 	p = strchr(path, '/');
 
 	for_each_child(tree, child) {
-		if (p && strneq(path, child->name, p-path))
+		if (p && (strlen(child->name) == p-path) &&
+				strneq(path, child->name, p-path))
 			return get_node_by_path(child, p+1);
 		else if (!p && streq(path, child->name))
 			return child;
diff --git a/tests/path-references.c b/tests/path-references.c
index c8d25fb..5e332e8 100644
--- a/tests/path-references.c
+++ b/tests/path-references.c
@@ -66,7 +66,7 @@ int main(int argc, char *argv[])
 	void *fdt;
 	const char *p;
 	int len, multilen;
-	int n1, n2;
+	int n1, n2, n3, n4;
 
 	test_init(argc, argv);
 	fdt = load_blob_arg(argc, argv);
@@ -92,6 +92,16 @@ int main(int argc, char *argv[])
 	if ((!streq(p, "/node1") || !streq(p + strlen("/node1") + 1, "/node2")))
 		FAIL("multiref has wrong value");
 
+	/* Check reference to nested nodes with common prefix */
+	n3 = fdt_path_offset(fdt, "/foo/baz");
+	if (n3 < 0)
+		FAIL("fdt_path_offset(/foo/baz): %s", fdt_strerror(n3));
+	n4 = fdt_path_offset(fdt, "/foobar/baz");
+	if (n4 < 0)
+		FAIL("fdt_path_offset(/foobar/baz): %s", fdt_strerror(n4));
+	check_ref(fdt, n3, "/foobar/baz");
+	check_ref(fdt, n4, "/foo/baz");
+
 	check_rref(fdt);
 
 	PASS();
diff --git a/tests/path-references.dts b/tests/path-references.dts
index b00fd79..8c66d80 100644
--- a/tests/path-references.dts
+++ b/tests/path-references.dts
@@ -12,4 +12,17 @@
 		ref = &{/node1}; /* reference after target */
 		lref = &n1;
 	};
+	/* Check references to nested nodes with common prefix */
+	foobar {
+		n3: baz {
+			ref = &{/foo/baz};
+			lref = &n4;
+		};
+	};
+	foo {
+		n4: baz {
+			ref = &{/foobar/baz};
+			lref = &n3;
+		};
+	};
 };
-- 
2.7.4

Re: [PATCH] Fix get_node_by_path string equality check

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 1845 bytes --]

On Wed, Apr 19, 2017 at 01:15:08AM +0000, Tim Montague wrote:
>  > Ah.. it looks like your mailer has mangled the patch, I'm afraid:
> > 
> > $ git am -s ~/foobox/
> > Applying: Fix get_node_by_path string equality check
> > .git/rebase-apply/patch:16: trailing whitespace.
> > 		if (p && (strlen(child->name) == p-path) &&
> > .git/rebase-apply/patch:17: trailing whitespace.
> > 				strneq(path, child->name, p-path))
> > .git/rebase-apply/patch:30: trailing whitespace.
> > 	int n1, n2, n3, n4;
> > .git/rebase-apply/patch:38: trailing whitespace.
> > 	/* Check reference to nested nodes with common prefix */
> > .git/rebase-apply/patch:39: trailing whitespace.
> > 	n3 = fdt_path_offset(fdt, "/foo/baz");
> > error: patch failed: livetree.c:478
> > error: livetree.c: patch does not apply
> > error: patch failed: tests/path-references.c:66
> > error: tests/path-references.c: patch does not apply
> > error: patch failed: tests/path-references.dts:12
> > error: tests/path-references.dts: patch does not apply
> > Patch failed at 0001 Fix get_node_by_path string equality check
> > The copy of the patch that failed is found in: .git/rebase-apply/patch
> > When you have resolved this problem, run "git am --continue".
> > If you prefer to skip this patch, run "git am --skip" instead.
> > To restore the original branch and stop patching, run "git am --abort".
> > 
> > Can you check your mailer settings.  If it's not clear how to stop it
> > doing that, it's probably safe to attach the patch instead.
> 
> Sorry about that, Outlook doesn't always behave. The patch is attached.

Thanks, applied to master.

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 819 bytes --]