strncpy() is deprecated for use on NUL-terminated destination strings
[1] and as such we should prefer more robust and less ambiguous string
interfaces.
The previous code took special care of NUL-terminating the destination
buffer via a manual assignment. As such, there is no bug in the current
implementation. However, in an effort to rid the kernel of strscpy()
[2], Let's instead use strscpy() which guarantees this behavior [3]. To
ensure we can copy the same number of bytes as before, add 1 to the
length argument provided to strscpy().
Link: https://www.kernel.org/doc/html/latest/process/deprecated.html#strncpy-on-nul-terminated-strings [1]
Link: https://github.com/KSPP/linux/issues/90 [2]
Link: https://manpages.debian.org/testing/linux-manual-4.8/strscpy.9.en.html [3]
Cc: [email protected]
Signed-off-by: Justin Stitt <[email protected]>
---
Note: build-tested only.
Found with: $ rg "strncpy\("
---
fs/squashfs/namei.c | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/fs/squashfs/namei.c b/fs/squashfs/namei.c
index 11e4539b9eae..6c4704ba8f42 100644
--- a/fs/squashfs/namei.c
+++ b/fs/squashfs/namei.c
@@ -80,8 +80,7 @@ static int get_dir_index_using_name(struct super_block *sb,
}
str = &index->name[SQUASHFS_NAME_LEN + 1];
- strncpy(str, name, len);
- str[len] = '\0';
+ strscpy(str, name, len + 1);
for (i = 0; i < i_count; i++) {
err = squashfs_read_metadata(sb, index, &index_start,
---
base-commit: 928a87efa42302a23bb9554be081a28058495f22
change-id: 20240328-strncpy-fs-squashfs-namei-c-9d01b8975e53
Best regards,
--
Justin Stitt <[email protected]>
On Thu, Mar 28, 2024 at 09:52:59PM +0000, Justin Stitt wrote:
> strncpy() is deprecated for use on NUL-terminated destination strings
> [1] and as such we should prefer more robust and less ambiguous string
> interfaces.
>
> The previous code took special care of NUL-terminating the destination
> buffer via a manual assignment. As such, there is no bug in the current
> implementation. However, in an effort to rid the kernel of strscpy()
typo: rid kernel of strncpy. :)
> [2], Let's instead use strscpy() which guarantees this behavior [3]. To
> ensure we can copy the same number of bytes as before, add 1 to the
> length argument provided to strscpy().
>
> Link: https://www.kernel.org/doc/html/latest/process/deprecated.html#strncpy-on-nul-terminated-strings [1]
> Link: https://github.com/KSPP/linux/issues/90 [2]
> Link: https://manpages.debian.org/testing/linux-manual-4.8/strscpy.9.en.html [3]
> Cc: [email protected]
> Signed-off-by: Justin Stitt <[email protected]>
> ---
> Note: build-tested only.
>
> Found with: $ rg "strncpy\("
> ---
> fs/squashfs/namei.c | 3 +--
> 1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/fs/squashfs/namei.c b/fs/squashfs/namei.c
> index 11e4539b9eae..6c4704ba8f42 100644
> --- a/fs/squashfs/namei.c
> +++ b/fs/squashfs/namei.c
> @@ -80,8 +80,7 @@ static int get_dir_index_using_name(struct super_block *sb,
> }
>
> str = &index->name[SQUASHFS_NAME_LEN + 1];
> - strncpy(str, name, len);
> - str[len] = '\0';
> + strscpy(str, name, len + 1);
Otherwise, yeah, looks right.
Reviewed-by: Kees Cook <[email protected]>
-Kees
>
> for (i = 0; i < i_count; i++) {
> err = squashfs_read_metadata(sb, index, &index_start,
>
> ---
> base-commit: 928a87efa42302a23bb9554be081a28058495f22
> change-id: 20240328-strncpy-fs-squashfs-namei-c-9d01b8975e53
>
> Best regards,
> --
> Justin Stitt <[email protected]>
>
>
--
Kees Cook
On 28/03/2024 21:52, Justin Stitt wrote:
> strncpy() is deprecated for use on NUL-terminated destination strings
> [1] and as such we should prefer more robust and less ambiguous string
> interfaces.
>
> The previous code took special care of NUL-terminating the destination
> buffer via a manual assignment. As such, there is no bug in the current
> implementation. However, in an effort to rid the kernel of strscpy()
> [2], Let's instead use strscpy() which guarantees this behavior [3]. To
> ensure we can copy the same number of bytes as before, add 1 to the
> length argument provided to strscpy().
>
Squashfs copies the passed string into a temporary buffer to ensure it
is NUL-terminated. This however is completely unnecessary as the
string is already NUL-terminated.
A better way to remove the strncpy() is to remove the unnecessary string
copy, which I have done in this patch here:
https://lore.kernel.org/lkml/[email protected]/T/#u
Phillip
> Link: https://www.kernel.org/doc/html/latest/process/deprecated.html#strncpy-on-nul-terminated-strings [1]
> Link: https://github.com/KSPP/linux/issues/90 [2]
> Link: https://manpages.debian.org/testing/linux-manual-4.8/strscpy.9.en.html [3]
> Cc: [email protected]
> Signed-off-by: Justin Stitt <[email protected]>
> ---
> Note: build-tested only.
>
> Found with: $ rg "strncpy\("
> ---
> fs/squashfs/namei.c | 3 +--
> 1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/fs/squashfs/namei.c b/fs/squashfs/namei.c
> index 11e4539b9eae..6c4704ba8f42 100644
> --- a/fs/squashfs/namei.c
> +++ b/fs/squashfs/namei.c
> @@ -80,8 +80,7 @@ static int get_dir_index_using_name(struct super_block *sb,
> }
>
> str = &index->name[SQUASHFS_NAME_LEN + 1];
> - strncpy(str, name, len);
> - str[len] = '\0';
> + strscpy(str, name, len + 1);
>
> for (i = 0; i < i_count; i++) {
> err = squashfs_read_metadata(sb, index, &index_start,
>
> ---
> base-commit: 928a87efa42302a23bb9554be081a28058495f22
> change-id: 20240328-strncpy-fs-squashfs-namei-c-9d01b8975e53
>
> Best regards,
> --
> Justin Stitt <[email protected]>
>
On Wed, Apr 3, 2024 at 12:32 PM Phillip Lougher <[email protected]> wrote:
> A better way to remove the strncpy() is to remove the unnecessary string
> copy, which I have done in this patch here:
Great! Cleaning up this code while removing strncpy() is a two for one.
On 04/04/2024 01:30, Justin Stitt wrote:
> On Wed, Apr 3, 2024 at 12:32 PM Phillip Lougher <[email protected]> wrote:
>> A better way to remove the strncpy() is to remove the unnecessary string
>> copy, which I have done in this patch here:
>
> Great! Cleaning up this code while removing strncpy() is a two for one.
Yes - thanks for your reply.
Phillip