diff mbox series

[v2] hfsplus: refactor copy_name to not use strncpy

Message ID 20240401-strncpy-fs-hfsplus-xattr-c-v2-1-6e089999355e@google.com (mailing list archive)
State New
Headers show
Series [v2] hfsplus: refactor copy_name to not use strncpy | expand

Commit Message

Justin Stitt April 1, 2024, 6:10 p.m. UTC
strncpy() is deprecated with NUL-terminated destination strings [1].

The copy_name() method does a lot of manual buffer manipulation to
eventually arrive with its desired string. If we don't know the
namespace this attr has or belongs to we want to prepend "osx." to our
final string. Following this, we're copying xattr_name and doing a
bizarre manual NUL-byte assignment with a memset where n=1.

Really, we can use some more obvious string APIs to acomplish this,
improving readability and security. Following the same control flow as
before: if we don't know the namespace let's use scnprintf() to form our
prefix + xattr_name pairing (while NUL-terminating too!). Otherwise, use
strscpy() to return the number of bytes copied into our buffer.
Additionally, for non-empty strings, include the NUL-byte in the length
-- matching the behavior of the previous implementation.

Note that strscpy() _can_ return -E2BIG but this is already handled by
all callsites:

In both hfsplus_listxattr_finder_info() and hfsplus_listxattr(), ret is
already type ssize_t so we can change the return type of copy_name() to
match (understanding that scnprintf()'s return type is different yet
fully representable by ssize_t). Furthermore, listxattr() in fs/xattr.c
is well-equipped to handle a potential -E2BIG return result from
vfs_listxattr():
|	ssize_t error;
...
|	error = vfs_listxattr(d, klist, size);
|	if (error > 0) {
|		if (size && copy_to_user(list, klist, error))
|			error = -EFAULT;
|	} else if (error == -ERANGE && size >= XATTR_LIST_MAX) {
|		/* The file system tried to returned a list bigger
|			than XATTR_LIST_MAX bytes. Not possible. */
|		error = -E2BIG;
|	}
... the error can potentially already be -E2BIG, skipping this else-if
and ending up at the same state as other errors.

Link: https://www.kernel.org/doc/html/latest/process/deprecated.html#strncpy-on-nul-terminated-strings [1]
Link: https://manpages.debian.org/testing/linux-manual-4.8/strscpy.9.en.html
Link: https://github.com/KSPP/linux/issues/90
Cc: linux-hardening@vger.kernel.org
Signed-off-by: Justin Stitt <justinstitt@google.com>
---
Changes in v2:
- include NUL-byte in length (thanks Kees)
- reword commit message slightly
- Link to v1: https://lore.kernel.org/r/20240321-strncpy-fs-hfsplus-xattr-c-v1-1-0c6385a10251@google.com
---
---
 fs/hfsplus/xattr.c | 22 ++++++++++------------
 1 file changed, 10 insertions(+), 12 deletions(-)


---
base-commit: 241590e5a1d1b6219c8d3045c167f2fbcc076cbb
change-id: 20240321-strncpy-fs-hfsplus-xattr-c-4ebfe67f4c6d

Best regards,
--
Justin Stitt <justinstitt@google.com>

Comments

Kees Cook April 4, 2024, 9:25 p.m. UTC | #1
On Mon, Apr 01, 2024 at 06:10:48PM +0000, Justin Stitt wrote:
> strncpy() is deprecated with NUL-terminated destination strings [1].
> 
> The copy_name() method does a lot of manual buffer manipulation to
> eventually arrive with its desired string. If we don't know the
> namespace this attr has or belongs to we want to prepend "osx." to our
> final string. Following this, we're copying xattr_name and doing a
> bizarre manual NUL-byte assignment with a memset where n=1.
> 
> Really, we can use some more obvious string APIs to acomplish this,
> improving readability and security. Following the same control flow as
> before: if we don't know the namespace let's use scnprintf() to form our
> prefix + xattr_name pairing (while NUL-terminating too!). Otherwise, use
> strscpy() to return the number of bytes copied into our buffer.
> Additionally, for non-empty strings, include the NUL-byte in the length
> -- matching the behavior of the previous implementation.
> 
> Note that strscpy() _can_ return -E2BIG but this is already handled by
> all callsites:
> 
> In both hfsplus_listxattr_finder_info() and hfsplus_listxattr(), ret is
> already type ssize_t so we can change the return type of copy_name() to
> match (understanding that scnprintf()'s return type is different yet
> fully representable by ssize_t). Furthermore, listxattr() in fs/xattr.c
> is well-equipped to handle a potential -E2BIG return result from
> vfs_listxattr():
> |	ssize_t error;
> ...
> |	error = vfs_listxattr(d, klist, size);
> |	if (error > 0) {
> |		if (size && copy_to_user(list, klist, error))
> |			error = -EFAULT;
> |	} else if (error == -ERANGE && size >= XATTR_LIST_MAX) {
> |		/* The file system tried to returned a list bigger
> |			than XATTR_LIST_MAX bytes. Not possible. */
> |		error = -E2BIG;
> |	}
> ... the error can potentially already be -E2BIG, skipping this else-if
> and ending up at the same state as other errors.
> 
> Link: https://www.kernel.org/doc/html/latest/process/deprecated.html#strncpy-on-nul-terminated-strings [1]
> Link: https://manpages.debian.org/testing/linux-manual-4.8/strscpy.9.en.html
> Link: https://github.com/KSPP/linux/issues/90
> Cc: linux-hardening@vger.kernel.org
> Signed-off-by: Justin Stitt <justinstitt@google.com>

Thanks, this looks right to me now!

Reviewed-by: Kees Cook <keescook@chromium.org>
Kees Cook April 24, 2024, 11:55 p.m. UTC | #2
On Mon, 01 Apr 2024 18:10:48 +0000, Justin Stitt wrote:
> strncpy() is deprecated with NUL-terminated destination strings [1].
> 
> The copy_name() method does a lot of manual buffer manipulation to
> eventually arrive with its desired string. If we don't know the
> namespace this attr has or belongs to we want to prepend "osx." to our
> final string. Following this, we're copying xattr_name and doing a
> bizarre manual NUL-byte assignment with a memset where n=1.
> 
> [...]

Applied to for-next/hardening, thanks!

[1/1] hfsplus: refactor copy_name to not use strncpy
      https://git.kernel.org/kees/c/628a89b6ef79

Take care,
diff mbox series

Patch

diff --git a/fs/hfsplus/xattr.c b/fs/hfsplus/xattr.c
index 9c9ff6b8c6f7..5a400259ae74 100644
--- a/fs/hfsplus/xattr.c
+++ b/fs/hfsplus/xattr.c
@@ -400,21 +400,19 @@  static int name_len(const char *xattr_name, int xattr_name_len)
 	return len;
 }
 
-static int copy_name(char *buffer, const char *xattr_name, int name_len)
+static ssize_t copy_name(char *buffer, const char *xattr_name, int name_len)
 {
-	int len = name_len;
-	int offset = 0;
+	ssize_t len;
 
-	if (!is_known_namespace(xattr_name)) {
-		memcpy(buffer, XATTR_MAC_OSX_PREFIX, XATTR_MAC_OSX_PREFIX_LEN);
-		offset += XATTR_MAC_OSX_PREFIX_LEN;
-		len += XATTR_MAC_OSX_PREFIX_LEN;
-	}
-
-	strncpy(buffer + offset, xattr_name, name_len);
-	memset(buffer + offset + name_len, 0, 1);
-	len += 1;
+	if (!is_known_namespace(xattr_name))
+		len = scnprintf(buffer, name_len + XATTR_MAC_OSX_PREFIX_LEN,
+				 "%s%s", XATTR_MAC_OSX_PREFIX, xattr_name);
+	else
+		len = strscpy(buffer, xattr_name, name_len + 1);
 
+	/* include NUL-byte in length for non-empty name */
+	if (len >= 0)
+		len++;
 	return len;
 }