Linux NFS development
 help / color / mirror / Atom feed
From: Olga Kornievskaia <okorniev@redhat.com>
To: steved@redhat.com
Cc: linux-nfs@vger.kernel.org
Subject: [PATCH 2/3] libtirpc: limit XDR decode node count
Date: Sun,  6 Sep 2026 21:13:27 -0400	[thread overview]
Message-ID: <20260907011328.21425-3-okorniev@redhat.com> (raw)
In-Reply-To: <20260907011328.21425-1-okorniev@redhat.com>

xdr_pmaplist(), xdr_rpcblist_ptr() and xdr_rpcb_entry_list_ptr()
decode linked lists with no upper bound on element count. A malicious
server can stream arbitrarily many list nodes, forcing unbounded heap
allocation in any client that decodes the response (rpcinfo,
rpcb_getmaps() callers).

Signed-off-by: Olga Kornievskaia <okorniev@redhat.com>
---
 src/pmap_prot2.c |  9 ++++++++-
 src/rpcb_prot.c  | 16 ++++++++++++++++
 2 files changed, 24 insertions(+), 1 deletion(-)

diff --git a/src/pmap_prot2.c b/src/pmap_prot2.c
index 7611bd5..d043f34 100644
--- a/src/pmap_prot2.c
+++ b/src/pmap_prot2.c
@@ -39,6 +39,8 @@
 #include <rpc/xdr.h>
 #include <rpc/pmap_prot.h>
 
+/* Cap decoded list length to prevent memory exhaustion from malicious data */
+#define PMAPLIST_DECODE_MAX_NODES 1024
 
 /*
  * What is going on with linked lists? (!)
@@ -87,7 +89,7 @@ xdr_pmaplist(XDR *xdrs, struct pmaplist **rp)
 	 * xdr_bool when the direction is XDR_DECODE.
 	 */
 	bool_t more_elements;
-	int freeing;
+	int freeing, node_count = 0;
 	struct pmaplist *next = NULL;
 	struct pmaplist *next_copy;
 
@@ -102,6 +104,11 @@ xdr_pmaplist(XDR *xdrs, struct pmaplist **rp)
 			return (FALSE);
 		if (! more_elements)
 			return (TRUE);  /* we are done */
+		if (xdrs->x_op == XDR_DECODE &&
+			++node_count > PMAPLIST_DECODE_MAX_NODES) {
+			return (FALSE);
+		}
+
 		/*
 		 * the unfortunate side effect of non-recursion is that in
 		 * the case of freeing we must remember the next object
diff --git a/src/rpcb_prot.c b/src/rpcb_prot.c
index 809feb0..0b0574f 100644
--- a/src/rpcb_prot.c
+++ b/src/rpcb_prot.c
@@ -89,6 +89,9 @@ xdr_rpcb(
  * serialize the rpcb elements.
  */
 
+/* Cap decoded list length to prevent memory exhaustion from malicious data */
+#define RPCBLIST_DECODE_MAX_NODES 1024
+
 bool_t
 xdr_rpcblist_ptr(
 	XDR *xdrs,
@@ -101,6 +104,7 @@ xdr_rpcblist_ptr(
 	 */
 	bool_t more_elements;
 	int freeing = (xdrs->x_op == XDR_FREE);
+	int node_count = 0;
 	rpcblist_ptr next;
 	rpcblist_ptr next_copy;
 
@@ -113,6 +117,10 @@ xdr_rpcblist_ptr(
 		if (! more_elements) {
 			return (TRUE);  /* we are done */
 		}
+		if (xdrs->x_op == XDR_DECODE &&
+		    ++node_count > RPCBLIST_DECODE_MAX_NODES) {
+			return (FALSE);
+		}
 		/*
 		 * the unfortunate side effect of non-recursion is that in
 		 * the case of freeing we must remember the next object
@@ -178,6 +186,9 @@ xdr_rpcb_entry(
 	return (TRUE);
 }
 
+/* Cap decoded list length to prevent memory exhaustion from malicious data */
+#define RPCB_ENTRY_LIST_DECODE_MAX_NODES 1024
+
 bool_t
 xdr_rpcb_entry_list_ptr(
 	XDR *xdrs,
@@ -190,6 +201,7 @@ xdr_rpcb_entry_list_ptr(
 	 */
 	bool_t more_elements;
 	int freeing = (xdrs->x_op == XDR_FREE);
+	int node_count = 0;
 	rpcb_entry_list_ptr next;
 	rpcb_entry_list_ptr next_copy;
 
@@ -202,6 +214,10 @@ xdr_rpcb_entry_list_ptr(
 		if (! more_elements) {
 			return (TRUE);  /* we are done */
 		}
+		if (xdrs->x_op == XDR_DECODE &&
+		    ++node_count > RPCB_ENTRY_LIST_DECODE_MAX_NODES) {
+			return (FALSE);
+		}
 		/*
 		 * the unfortunate side effect of non-recursion is that in
 		 * the case of freeing we must remember the next object
-- 
2.52.0


  parent reply	other threads:[~2026-09-07  1:13 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-09-07  1:13 [PATCH 1/1] fsidd: require root credentials on abstract socket Olga Kornievskaia
2026-09-07  1:13 ` [PATCH 1/3] libtirpc: Fix use-after-free in xdr_pmaplist() XDR_FREE path Olga Kornievskaia
2026-09-07  1:13 ` Olga Kornievskaia [this message]
2026-09-07  1:13 ` [PATCH 3/3] libtirpc: Bound maxsize in xdr_rpc_gss_unwrap_data() decode calls Olga Kornievskaia

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260907011328.21425-3-okorniev@redhat.com \
    --to=okorniev@redhat.com \
    --cc=linux-nfs@vger.kernel.org \
    --cc=steved@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox