GNU bug report logs - #53928
[PATCH] Allow null bytes in UNIX sockets.

Previous Next

Package: guile;

Reported by: Liliana Marie Prikler <liliana.prikler <at> gmail.com>

Date: Thu, 10 Feb 2022 20:50:02 UTC

Severity: normal

Tags: patch

Done: Ludovic Courtès <ludo <at> gnu.org>

Bug is archived. No further changes may be made.

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 53928 in the body.
You can then email your comments to 53928 AT debbugs.gnu.org in the normal way.

Toggle the display of automated, internal messages from the tracker.

View this report as an mbox folder, status mbox, maintainer mbox


Report forwarded to bug-guile <at> gnu.org:
bug#53928; Package guile. (Thu, 10 Feb 2022 20:50:02 GMT) Full text and rfc822 format available.

Acknowledgement sent to Liliana Marie Prikler <liliana.prikler <at> gmail.com>:
New bug report received and forwarded. Copy sent to bug-guile <at> gnu.org. (Thu, 10 Feb 2022 20:50:02 GMT) Full text and rfc822 format available.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Liliana Marie Prikler <liliana.prikler <at> gmail.com>
To: bug-guile <at> gnu.org
Subject: [PATCH] Allow null bytes in UNIX sockets.
Date: Thu, 10 Feb 2022 21:42:31 +0100
The current socket address constructors all assume, that there are no
null bytes in the socket path.  This assumption does not hold in Linux,
which uses an initial null byte to demarcate abstract sockets and
ignores all further null bytes [1].

[1] https://www.man7.org/linux/man-pages/man7/unix.7.html

* libguile/sockets.c (scm_fill_sockaddr)[HAVE_UNIX_DOMAIN_SOCKETS]:
Use scm_to_locale_stringn to construct c_address.
Use memcpy instead of strcpy and calculate size directly instead of
using SUN_LEN.
(_scm_from_sockaddr): Copy the entire path up to the limits imposed by
addr_size.
* test-suite/tests/00-socket.test: ("make-socket-address"): Add case for
abstract unix sockets.
("AF_UNIX/SOCK_STREAM"): Add abstract socket versions of bind, listen,
connect and accept.
---
This patch is identical to the one sent in [1].  The full thread is visible
over at yhetil[2] -- one reply from May 2021 [3] isn't shown as part of it in
the GNU archive.

I haven't built or tested it on top of current master, but given that the
patch applied cleanly, I have high hopes to say the least 🙂

[1] https://lists.gnu.org/archive/html/guile-devel/2021-03/msg00038.html
[2] https://yhetil.org/guile-devel/20210329153757.16476-1-leo.priker <at> student.tugraz.at/
[3] https://lists.gnu.org/archive/html/guile-devel/2021-05/msg00012.html

 libguile/socket.c               | 18 +++++---
 test-suite/tests/00-socket.test | 78 +++++++++++++++++++++++++++++++--
 2 files changed, 86 insertions(+), 10 deletions(-)

diff --git a/libguile/socket.c b/libguile/socket.c
index 77cdd74ef..6b77b3a14 100644
--- a/libguile/socket.c
+++ b/libguile/socket.c
@@ -813,10 +813,11 @@ scm_fill_sockaddr (int fam, SCM address, SCM *args, int which_arg,
 	struct sockaddr_un *soka;
 	int addr_size;
 	char *c_address;
+	size_t c_address_size;
 
 	scm_dynwind_begin (0);
 
-	c_address = scm_to_locale_string (address);
+	c_address = scm_to_locale_stringn (address, &c_address_size);
 	scm_dynwind_free (c_address);
 
 	/* the static buffer size in sockaddr_un seems to be arbitrary
@@ -826,12 +827,14 @@ scm_fill_sockaddr (int fam, SCM address, SCM *args, int which_arg,
 	   connect/bind etc., to fail.  sun_path is always the last
 	   member of the structure.  */
 	addr_size = sizeof (struct sockaddr_un)
-	  + MAX (0, strlen (c_address) + 1 - (sizeof soka->sun_path));
+	  + MAX (0, c_address_size + 1 - (sizeof soka->sun_path));
 	soka = (struct sockaddr_un *) scm_malloc (addr_size);
-	memset (soka, 0, addr_size);  /* for sun_len: see sin_len above. */
+	memset (soka, 0, addr_size);
 	soka->sun_family = AF_UNIX;
-	strcpy (soka->sun_path, c_address);
-	*size = SUN_LEN (soka);
+        /* we accept 0-bytes here (used for abstract sockets in Linux);
+           therefore do not use strlen() or SUN_LEN!  */
+	memcpy (soka->sun_path, c_address, c_address_size);
+	*size = offsetof (struct sockaddr_un, sun_path) + c_address_size;
 
 	scm_dynwind_end ();
 	return (struct sockaddr *) soka;
@@ -1045,7 +1048,10 @@ _scm_from_sockaddr (const scm_t_max_sockaddr *address, unsigned addr_size,
 	if (addr_size <= offsetof (struct sockaddr_un, sun_path))
 	  SCM_SIMPLE_VECTOR_SET(result, 1, SCM_BOOL_F);
 	else
-	  SCM_SIMPLE_VECTOR_SET(result, 1, scm_from_locale_string (nad->sun_path));
+          {
+            size_t path_size = addr_size - offsetof (struct sockaddr_un, sun_path);
+            SCM_SIMPLE_VECTOR_SET(result, 1, scm_from_locale_stringn(nad->sun_path, path_size));
+          }
       }
       break;
 #endif
diff --git a/test-suite/tests/00-socket.test b/test-suite/tests/00-socket.test
index 027bd8519..5196b4b7d 100644
--- a/test-suite/tests/00-socket.test
+++ b/test-suite/tests/00-socket.test
@@ -128,10 +128,15 @@
 	       (= (sockaddr:flowinfo sa*) 1)))))
 
   (if (defined? 'AF_UNIX)
-      (pass-if "AF_UNIX"
-	(let ((sa (make-socket-address AF_UNIX "/tmp/unix-socket")))
-	  (and (= (sockaddr:fam sa) AF_UNIX)
-	       (string=? (sockaddr:path sa) "/tmp/unix-socket"))))))
+      (begin
+        (pass-if "AF_UNIX"
+	  (let ((sa (make-socket-address AF_UNIX "/tmp/unix-socket")))
+	    (and (= (sockaddr:fam sa) AF_UNIX)
+	         (string=? (sockaddr:path sa) "/tmp/unix-socket"))))
+        (pass-if "AF_UNIX abstract"
+          (let ((sa (make-socket-address AF_UNIX "\x00/tmp/abstract-socket")))
+	    (and (= (sockaddr:fam sa) AF_UNIX)
+	         (string=? (sockaddr:path sa) "\x00/tmp/abstract-socket")))))))
 
 ;;;
 ;;; setsockopt
@@ -319,6 +324,71 @@
 
 	#t)
 
+      ;; testing `bind', `listen' and `connect' on abstract stream-oriented sockets
+
+      (let ((server-socket (socket AF_UNIX SOCK_STREAM 0))
+	    (server-bound? #f)
+	    (server-listening? #f)
+	    (server-pid #f)
+	    (path (temp-file-path)))
+
+	(false-if-exception (delete-file path))
+        (set! path (string-append "\x00" path))
+
+	(pass-if "bind abstract"
+	  (catch 'system-error
+	    (lambda ()
+	      (bind server-socket AF_UNIX path)
+	      (set! server-bound? #t)
+	      #t)
+	    (lambda args
+	      (let ((errno (system-error-errno args)))
+		(cond ((= errno EADDRINUSE) (throw 'unresolved))
+		      (else (apply throw args)))))))
+
+	(pass-if "listen abstract"
+	  (if (not server-bound?)
+	      (throw 'unresolved)
+	      (begin
+		(listen server-socket 123)
+		(set! server-listening? #t)
+		#t)))
+
+	(force-output (current-output-port))
+	(force-output (current-error-port))
+	(when server-listening?
+          (let ((pid (primitive-fork-if-available)))
+	    ;; Spawn a server process.
+	    (case pid
+	      ((-1)  ;; fork not available
+               #f)
+	      ((0)   ;; the kid:  serve one connection and exit
+	       (let serve ((conn
+			    (false-if-exception (accept server-socket)))
+			   (count 0))
+		 (if (not conn)
+		     (exit 1)
+	             (exit 0))))
+	      (else  ;; the parent
+	       (set! server-pid pid)
+	       #t))))
+
+	(pass-if "connect abstract"
+	  (if (not server-pid)
+	      (throw 'unresolved)
+	      (let ((s (socket AF_UNIX SOCK_STREAM 0)))
+                (display "connect abstract\n")
+		(connect s AF_UNIX path)
+		#t)))
+
+	(pass-if "accept abstract"
+	  (if (not server-pid)
+	      (throw 'unresolved)
+              (begin
+	        (let ((status (cdr (waitpid server-pid))))
+		  (eqv? 0 (status:exit-val status))))))
+
+	#t)
 
       ;; Testing `send', `recv!' & co. on stream-oriented sockets (with
       ;; a bit of duplication with the above.)
-- 
2.34.0





Reply sent to Ludovic Courtès <ludo <at> gnu.org>:
You have taken responsibility. (Thu, 16 Jun 2022 08:10:02 GMT) Full text and rfc822 format available.

Notification sent to Liliana Marie Prikler <liliana.prikler <at> gmail.com>:
bug acknowledged by developer. (Thu, 16 Jun 2022 08:10:02 GMT) Full text and rfc822 format available.

Message #10 received at 53928-done <at> debbugs.gnu.org (full text, mbox):

From: Ludovic Courtès <ludo <at> gnu.org>
To: Liliana Marie Prikler <liliana.prikler <at> gmail.com>
Cc: 53928-done <at> debbugs.gnu.org
Subject: Re: bug#53928: [PATCH] Allow null bytes in UNIX sockets.
Date: Thu, 16 Jun 2022 10:09:13 +0200
Hi,

Liliana Marie Prikler <liliana.prikler <at> gmail.com> skribis:

> The current socket address constructors all assume, that there are no
> null bytes in the socket path.  This assumption does not hold in Linux,
> which uses an initial null byte to demarcate abstract sockets and
> ignores all further null bytes [1].
>
> [1] https://www.man7.org/linux/man-pages/man7/unix.7.html
>
> * libguile/sockets.c (scm_fill_sockaddr)[HAVE_UNIX_DOMAIN_SOCKETS]:
> Use scm_to_locale_stringn to construct c_address.
> Use memcpy instead of strcpy and calculate size directly instead of
> using SUN_LEN.
> (_scm_from_sockaddr): Copy the entire path up to the limits imposed by
> addr_size.
> * test-suite/tests/00-socket.test: ("make-socket-address"): Add case for
> abstract unix sockets.
> ("AF_UNIX/SOCK_STREAM"): Add abstract socket versions of bind, listen,
> connect and accept.

Finally applied, thanks!

Ludo’.




bug archived. Request was from Debbugs Internal Request <help-debbugs <at> gnu.org> to internal_control <at> debbugs.gnu.org. (Thu, 14 Jul 2022 11:24:08 GMT) Full text and rfc822 format available.

This bug report was last modified 1 year and 281 days ago.

Previous Next


GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.