From: Doran Moppert <dmoppert@redhat.com>
To: linux-nfs@vger.kernel.org
Subject: [PATCH] rpcbind: pair all svc_getargs() calls with svc_freeargs() to avoid memory leak
Date: Thu, 11 May 2017 15:35:20 +0930 [thread overview]
Message-ID: <20170511060520.GE3866@sin.redhat.com> (raw)
This patch is to address CVE-2017-8779 "rpcbomb" in rpcbind, discussed
at [1], [2], [3]. The last link suggests this issue is actually a bug
in rpcbind, which led me here.
The leak caused by the reproducer at [4] appears to come from
rpcb_service_4(), in the case where svc_getargs() returns false and the
function had an early return, rather than passing through the cleanup
path at done:, as would otherwise occur.
It also addresses a couple of other locations where the same fault seems
to exist, though I haven't been able to exercise those. I hope someone
more intimate with rpc(3) can confirm my understanding is correct, and
that I haven't introduced any new bugs.
Without this patch, using the reproducer (and variants) repeatedly
against rpcbind with a numBytes argument of 1_000_000_000, /proc/$(pidof
rpcbind)/status reports VmSize increase of 976564 kB each call, and
VmRSS increase of around 260 kB every 33 calls - the specific numbers
are probably an artifact of my rhel/glibc version. With the patch,
there is a small (~50 kB) VmSize increase with the first message, but
thereafter both VmSize and VmRSS remain steady.
[1]: http://seclists.org/oss-sec/2017/q2/209
[2]: https://bugzilla.redhat.com/show_bug.cgi?id=1448124
[3]: https://sourceware.org/ml/libc-alpha/2017-05/msg00129.html
[4]: https://github.com/guidovranken/rpcbomb/
Signed-off-by: Doran Moppert <dmoppert@redhat.com>
---
src/pmap_svc.c | 56 +++++++++++++++++++++++++++++++++++++++++++++---------
src/rpcb_svc.c | 2 +-
src/rpcb_svc_4.c | 2 +-
src/rpcb_svc_com.c | 8 ++++++++
4 files changed, 57 insertions(+), 11 deletions(-)
diff --git a/src/pmap_svc.c b/src/pmap_svc.c
index 4c744fe..e926cdc 100644
--- a/src/pmap_svc.c
+++ b/src/pmap_svc.c
@@ -175,6 +175,7 @@ pmapproc_change(struct svc_req *rqstp /*__unused*/, SVCXPRT *xprt, unsigned long
long ans;
uid_t uid;
char uidbuf[32];
+ int rc = TRUE;
/*
* Can't use getpwnam here. We might end up calling ourselves
@@ -194,7 +195,8 @@ pmapproc_change(struct svc_req *rqstp /*__unused*/, SVCXPRT *xprt, unsigned long
if (!svc_getargs(xprt, (xdrproc_t) xdr_pmap, (char *)®)) {
svcerr_decode(xprt);
- return (FALSE);
+ rc = FALSE;
+ goto done;
}
#ifdef RPCBIND_DEBUG
if (debugging)
@@ -205,7 +207,8 @@ pmapproc_change(struct svc_req *rqstp /*__unused*/, SVCXPRT *xprt, unsigned long
if (!check_access(xprt, op, reg.pm_prog, PMAPVERS)) {
svcerr_weakauth(xprt);
- return (FALSE);
+ rc = (FALSE);
+ goto done;
}
rpcbreg.r_prog = reg.pm_prog;
@@ -258,7 +261,16 @@ done_change:
rpcbs_set(RPCBVERS_2_STAT, ans);
else
rpcbs_unset(RPCBVERS_2_STAT, ans);
- return (TRUE);
+done:
+ if (!svc_freeargs(xprt, (xdrproc_t) xdr_pmap, (char *)®)) {
+ if (debugging) {
+ (void) xlog(LOG_DEBUG, "unable to free arguments\n");
+ if (doabort) {
+ rpcbind_abort();
+ }
+ }
+ }
+ return (rc);
}
/* ARGSUSED */
@@ -272,15 +284,18 @@ pmapproc_getport(struct svc_req *rqstp /*__unused*/, SVCXPRT *xprt)
#ifdef RPCBIND_DEBUG
char *uaddr;
#endif
+ int rc = TRUE;
if (!svc_getargs(xprt, (xdrproc_t) xdr_pmap, (char *)®)) {
svcerr_decode(xprt);
- return (FALSE);
+ rc = FALSE;
+ goto done;
}
if (!check_access(xprt, PMAPPROC_GETPORT, reg.pm_prog, PMAPVERS)) {
svcerr_weakauth(xprt);
- return FALSE;
+ rc = FALSE;
+ goto done;
}
#ifdef RPCBIND_DEBUG
@@ -330,21 +345,34 @@ pmapproc_getport(struct svc_req *rqstp /*__unused*/, SVCXPRT *xprt)
pmap_ipprot2netid(reg.pm_prot) ?: "<unknown>",
port ? udptrans : "");
- return (TRUE);
+done:
+ if (!svc_freeargs(xprt, (xdrproc_t) xdr_pmap, (char *)®)) {
+ if (debugging) {
+ (void) xlog(LOG_DEBUG, "unable to free arguments\n");
+ if (doabort) {
+ rpcbind_abort();
+ }
+ }
+ }
+ return (rc);
}
/* ARGSUSED */
static bool_t
pmapproc_dump(struct svc_req *rqstp /*__unused*/, SVCXPRT *xprt)
{
+ int rc = TRUE;
+
if (!svc_getargs(xprt, (xdrproc_t)xdr_void, NULL)) {
svcerr_decode(xprt);
- return (FALSE);
+ rc = FALSE;
+ goto done;
}
if (!check_access(xprt, PMAPPROC_DUMP, 0, PMAPVERS)) {
svcerr_weakauth(xprt);
- return FALSE;
+ rc = FALSE;
+ goto done;
}
if ((!svc_sendreply(xprt, (xdrproc_t) xdr_pmaplist_ptr,
@@ -354,7 +382,17 @@ pmapproc_dump(struct svc_req *rqstp /*__unused*/, SVCXPRT *xprt)
rpcbind_abort();
}
}
- return (TRUE);
+
+done:
+ if (!svc_freeargs(xprt, (xdrproc_t) xdr_pmap, (char *)NULL)) {
+ if (debugging) {
+ (void) xlog(LOG_DEBUG, "unable to free arguments\n");
+ if (doabort) {
+ rpcbind_abort();
+ }
+ }
+ }
+ return (rc);
}
int pmap_netid2ipprot(const char *netid)
diff --git a/src/rpcb_svc.c b/src/rpcb_svc.c
index 709e3fb..091f530 100644
--- a/src/rpcb_svc.c
+++ b/src/rpcb_svc.c
@@ -166,7 +166,7 @@ rpcb_service_3(struct svc_req *rqstp, SVCXPRT *transp)
svcerr_decode(transp);
if (debugging)
(void) xlog(LOG_DEBUG, "rpcbind: could not decode");
- return;
+ goto done;
}
if (rqstp->rq_proc == RPCBPROC_SET
diff --git a/src/rpcb_svc_4.c b/src/rpcb_svc_4.c
index 5094879..eebbbbe 100644
--- a/src/rpcb_svc_4.c
+++ b/src/rpcb_svc_4.c
@@ -218,7 +218,7 @@ rpcb_service_4(struct svc_req *rqstp, SVCXPRT *transp)
svcerr_decode(transp);
if (debugging)
(void) xlog(LOG_DEBUG, "rpcbind: could not decode\n");
- return;
+ goto done;
}
if (rqstp->rq_proc == RPCBPROC_SET
diff --git a/src/rpcb_svc_com.c b/src/rpcb_svc_com.c
index 5862c26..cb63afd 100644
--- a/src/rpcb_svc_com.c
+++ b/src/rpcb_svc_com.c
@@ -927,6 +927,14 @@ error:
if (call_msg.rm_xid != 0)
(void) free_slot_by_xid(call_msg.rm_xid);
out:
+ if (!svc_freeargs(transp, (xdrproc_t) xdr_rmtcall_args, (char *) &a)) {
+ if (debugging) {
+ (void) xlog(LOG_DEBUG, "unable to free arguments\n");
+ if (doabort) {
+ rpcbind_abort();
+ }
+ }
+ }
if (local_uaddr)
free(local_uaddr);
if (buf_alloc)
--
1.8.3.1
--
Doran Moppert
Red Hat Product Security
next reply other threads:[~2017-05-11 6:05 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-05-11 6:05 Doran Moppert [this message]
2017-05-11 15:49 ` [PATCH] rpcbind: pair all svc_getargs() calls with svc_freeargs() to avoid memory leak Steve Dickson
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=20170511060520.GE3866@sin.redhat.com \
--to=dmoppert@redhat.com \
--cc=linux-nfs@vger.kernel.org \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.