[PATCH RESEND] apparmor: Replace deprecated strcpy with memcpy in gen_symlink_name

[PATCH RESEND] apparmor: Replace deprecated strcpy with memcpy in gen_symlink_name

[Thorsten Blum]

strcpy() is deprecated; use memcpy() instead. Unlike strcpy(), memcpy()
does not copy the NUL terminator from the source string, which would be
overwritten anyway on every iteration when using strcpy(). snprintf()
then ensures that 'char *s' is NUL-terminated.

Replace the hard-coded path length to remove the magic number 6, and add
a comment explaining the extra 11 bytes.

Link: https://github.com/KSPP/linux/issues/88
Signed-off-by: Thorsten Blum <thorsten.blum@linux.dev>
---
 security/apparmor/apparmorfs.c | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/security/apparmor/apparmorfs.c b/security/apparmor/apparmorfs.c
index 391a586d0557..4b2752200ce2 100644
--- a/security/apparmor/apparmorfs.c
+++ b/security/apparmor/apparmorfs.c
@@ -1602,16 +1602,20 @@ static char *gen_symlink_name(int depth, const char *dirname, const char *fname)
 {
 	char *buffer, *s;
 	int error;
-	int size = depth * 6 + strlen(dirname) + strlen(fname) + 11;
+	const char *path = "../../";
+	size_t path_len = strlen(path);
+	int size;
 
+	/* Extra 11 bytes: "raw_data" (9) + two slashes "//" (2) */
+	size = depth * path_len + strlen(dirname) + strlen(fname) + 11;
 	s = buffer = kmalloc(size, GFP_KERNEL);
 	if (!buffer)
 		return ERR_PTR(-ENOMEM);
 
 	for (; depth > 0; depth--) {
-		strcpy(s, "../../");
-		s += 6;
-		size -= 6;
+		memcpy(s, path, path_len);
+		s += path_len;
+		size -= path_len;
 	}
 
 	error = snprintf(s, size, "raw_data/%s/%s", dirname, fname);
-- 
2.51.1

Re: [PATCH RESEND] apparmor: Replace deprecated strcpy with memcpy in gen_symlink_name

[John Johansen]

On 11/26/25 08:57, Thorsten Blum wrote:
> strcpy() is deprecated; use memcpy() instead. Unlike strcpy(), memcpy()
> does not copy the NUL terminator from the source string, which would be
> overwritten anyway on every iteration when using strcpy(). snprintf()
> then ensures that 'char *s' is NUL-terminated.
> 
> Replace the hard-coded path length to remove the magic number 6, and add
> a comment explaining the extra 11 bytes.
> 
> Link: https://github.com/KSPP/linux/issues/88
> Signed-off-by: Thorsten Blum <thorsten.blum@linux.dev>

hey Thorsten,

sorry I have actually pulled these in, and tested them. I didn't send out
the acks yet because I have another patch that I was waiting on a proper
signed-off-by: on.

I am going to have to pull that one so I can push. I'll add acks now but
the push isn't going to happen for a few hours.

Acked-by: John Johansen <john.johansen@canonical.com>

>   security/apparmor/apparmorfs.c | 12 ++++++++----
>   1 file changed, 8 insertions(+), 4 deletions(-)
> 
> diff --git a/security/apparmor/apparmorfs.c b/security/apparmor/apparmorfs.c
> index 391a586d0557..4b2752200ce2 100644
> --- a/security/apparmor/apparmorfs.c
> +++ b/security/apparmor/apparmorfs.c
> @@ -1602,16 +1602,20 @@ static char *gen_symlink_name(int depth, const char *dirname, const char *fname)
>   {
>   	char *buffer, *s;
>   	int error;
> -	int size = depth * 6 + strlen(dirname) + strlen(fname) + 11;
> +	const char *path = "../../";
> +	size_t path_len = strlen(path);
> +	int size;
>   
> +	/* Extra 11 bytes: "raw_data" (9) + two slashes "//" (2) */
> +	size = depth * path_len + strlen(dirname) + strlen(fname) + 11;
>   	s = buffer = kmalloc(size, GFP_KERNEL);
>   	if (!buffer)
>   		return ERR_PTR(-ENOMEM);
>   
>   	for (; depth > 0; depth--) {
> -		strcpy(s, "../../");
> -		s += 6;
> -		size -= 6;
> +		memcpy(s, path, path_len);
> +		s += path_len;
> +		size -= path_len;
>   	}
>   
>   	error = snprintf(s, size, "raw_data/%s/%s", dirname, fname);

Re: [PATCH RESEND] apparmor: Replace deprecated strcpy with memcpy in gen_symlink_name

[Thorsten Blum]

On 27. Nov 2025, at 02:32, John Johansen wrote:
> hey Thorsten,
> 
> sorry I have actually pulled these in, and tested them. I didn't send out
> the acks yet because I have another patch that I was waiting on a proper
> signed-off-by: on.
> 
> I am going to have to pull that one so I can push. I'll add acks now but
> the push isn't going to happen for a few hours.
> 
> Acked-by: John Johansen <john.johansen@canonical.com>

Ah sorry for resending then, didn't know you looked at them already.

Thanks,
Thorsten