diff mbox

CVE-2016-4429: sunrpc: Do not use alloca in clntudp_call [BZ #20112]

Message ID 20160519110545.ED146400FD12E@oldenburg.str.redhat.com
State New
Headers show

Commit Message

Florian Weimer May 19, 2016, 11:05 a.m. UTC
The call is technically in a loop, and under certain circumstances
(which are quite difficult to reproduce in a test case), alloca
can be invoked repeatedly during a single call to clntudp_call.
As a result, the available stack space can be exhausted (even
though individual alloca sizes are bounded implicitly by what
can fit into a UDP packet, as a side effect of the earlier
successful send operation).

2016-05-19  Florian Weimer  <fweimer@redhat.com>

	CVE-2016-4429
	[BZ #20112]
	* sunrpc/clnt_udp.c (clntudp_call): Use malloc/free for the error
	payload.

Comments

Andreas Schwab May 19, 2016, 11:53 a.m. UTC | #1
fweimer@redhat.com (Florian Weimer) writes:

> The call is technically in a loop, and under certain circumstances
> (which are quite difficult to reproduce in a test case), alloca
> can be invoked repeatedly during a single call to clntudp_call.
> As a result, the available stack space can be exhausted (even
> though individual alloca sizes are bounded implicitly by what
> can fit into a UDP packet, as a side effect of the earlier
> successful send operation).

If you use a VLA you can avoid that.

Andreas.
Florian Weimer May 19, 2016, 11:57 a.m. UTC | #2
On 05/19/2016 01:53 PM, Andreas Schwab wrote:
> fweimer@redhat.com (Florian Weimer) writes:
>
>> The call is technically in a loop, and under certain circumstances
>> (which are quite difficult to reproduce in a test case), alloca
>> can be invoked repeatedly during a single call to clntudp_call.
>> As a result, the available stack space can be exhausted (even
>> though individual alloca sizes are bounded implicitly by what
>> can fit into a UDP packet, as a side effect of the earlier
>> successful send operation).
>
> If you use a VLA you can avoid that.

It's still a maintenance hazard for libtirpc because they might 
eventually support IPv6 jumbograms, which won't fit on the stack.

Florian
Florian Weimer May 23, 2016, 8:57 a.m. UTC | #3
I want to commit the original patch later today.

Thanks,
Florian
diff mbox

Patch

diff --git a/NEWS b/NEWS
index b3fd3cc..2341697 100644
--- a/NEWS
+++ b/NEWS
@@ -48,6 +48,10 @@  Security related changes:
   called with the GLOB_ALTDIRFUNC flag and encountered a long file name.
   Reported by Alexander Cherepanov.  (CVE-2016-1234)
 
+* The Sun RPC UDP client could exhaust all available stack space when
+  flooded with crafted ICMP and UDP messages.  Reported by Aldy Hernandez'
+  alloca plugin for GCC.  (CVE-2016-4429)
+
 The following bugs are resolved with this release:
 
   [The release manager will add the list generated by
diff --git a/sunrpc/clnt_udp.c b/sunrpc/clnt_udp.c
index a6cf5f1..4d9acb1 100644
--- a/sunrpc/clnt_udp.c
+++ b/sunrpc/clnt_udp.c
@@ -388,9 +388,15 @@  send_again:
 	  struct sock_extended_err *e;
 	  struct sockaddr_in err_addr;
 	  struct iovec iov;
-	  char *cbuf = (char *) alloca (outlen + 256);
+	  char *cbuf = malloc (outlen + 256);
 	  int ret;
 
+	  if (cbuf == NULL)
+	    {
+	      cu->cu_error.re_errno = errno;
+	      return (cu->cu_error.re_status = RPC_CANTRECV);
+	    }
+
 	  iov.iov_base = cbuf + 256;
 	  iov.iov_len = outlen;
 	  msg.msg_name = (void *) &err_addr;
@@ -415,10 +421,12 @@  send_again:
 		 cmsg = CMSG_NXTHDR (&msg, cmsg))
 	      if (cmsg->cmsg_level == SOL_IP && cmsg->cmsg_type == IP_RECVERR)
 		{
+		  free (cbuf);
 		  e = (struct sock_extended_err *) CMSG_DATA(cmsg);
 		  cu->cu_error.re_errno = e->ee_errno;
 		  return (cu->cu_error.re_status = RPC_CANTRECV);
 		}
+	  free (cbuf);
 	}
 #endif
       do