Linux NFS development
 help / color / mirror / Atom feed
From: Chuck Lever <cel@kernel.org>
To: Trond Myklebust <trond.myklebust@hammerspace.com>,
	Anna Schumaker <anna@kernel.org>, NeilBrown <neil@brown.name>,
	Jeff Layton <jlayton@kernel.org>,
	Olga Kornievskaia <okorniev@redhat.com>,
	Dai Ngo <dai.ngo@oracle.com>, Tom Talpey <tom@talpey.com>
Cc: <linux-nfs@vger.kernel.org>, Chuck Lever <chuck.lever@oracle.com>
Subject: [PATCH 0/4] sunrpc: close length validation gaps in krb5p unwrap
Date: Sat, 23 May 2026 21:02:09 -0400	[thread overview]
Message-ID: <20260524010213.557424-1-cel@kernel.org> (raw)

From: Chuck Lever <chuck.lever@oracle.com>

Neither svcauth_gss_unwrap_priv() nor gss_unwrap_resp_priv()
enforces a minimum on the wire-supplied opaque length before
forwarding it to the krb5 unwrap core.  An authenticated krb5p
peer -- NFS client or server -- can send a token shorter than the
16-byte RFC 4121 header and reach gss_krb5_unwrap_v2() with a
length that drives pre-decrypt out-of-bounds reads, a
divide-by-zero in _rotate_left(), and a post-decrypt unsigned
underflow in the movelen computation that feeds a multi-gigabyte
memmove.

The fixes are layered so that neither the caller boundary nor
the unwrap core depends on the other for safety.  The caller-side
floor checks (patches 1, 2) are sufficient to close the attack,
but gss_krb5_unwrap_v2() must also be self-defending because it
is reachable through gss_unwrap() from any future caller that
might not enforce the same floor.  The xdr_buf_trim() clamp
(patch 3) is independent of krb5 -- it corrects a generic
invariant violation that gss_krb5_unwrap_v2() happens to be the
only current trigger for, but that any caller could hit if
buf->len and the component iov_lens get out of sync.

The client side (patch 2) has an additional u32 integer-overflow
bypass that the server side does not: the original
`offset + opaque_len > rcv_buf->len` wraps when opaque_len is
near UINT_MAX, accepting a huge token length that re-enters the
same underflow chain.  The replacement splits the check into
separate offset and length guards that are safe in u32 arithmetic.

Chris Mason (4):
  SUNRPC: svcauth_gss: enforce krb5 token minimum length
  SUNRPC: harden gss_unwrap_resp_priv length checks
  SUNRPC: xdr_buf_trim: clamp buf->len to avoid underflow
  SUNRPC: harden gss_krb5_unwrap_v2 against short tokens

 net/sunrpc/auth_gss/auth_gss.c      |  6 +++++-
 net/sunrpc/auth_gss/gss_krb5_wrap.c | 11 +++++++++--
 net/sunrpc/auth_gss/svcauth_gss.c   |  2 ++
 net/sunrpc/xdr.c                    |  2 +-
 4 files changed, 17 insertions(+), 4 deletions(-)

-- 
2.54.0


             reply	other threads:[~2026-05-24  1:02 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-05-24  1:02 Chuck Lever [this message]
2026-05-24  1:02 ` [PATCH 1/4] SUNRPC: svcauth_gss: enforce krb5 token minimum length Chuck Lever
2026-05-24  1:02 ` [PATCH 2/4] SUNRPC: harden gss_unwrap_resp_priv length checks Chuck Lever
2026-05-24  1:02 ` [PATCH 3/4] SUNRPC: xdr_buf_trim: clamp buf->len to avoid underflow Chuck Lever
2026-05-24  1:02 ` [PATCH 4/4] SUNRPC: harden gss_krb5_unwrap_v2 against short tokens Chuck Lever
2026-05-24 10:56 ` [PATCH 0/4] sunrpc: close length validation gaps in krb5p unwrap Jeff Layton

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=20260524010213.557424-1-cel@kernel.org \
    --to=cel@kernel.org \
    --cc=anna@kernel.org \
    --cc=chuck.lever@oracle.com \
    --cc=dai.ngo@oracle.com \
    --cc=jlayton@kernel.org \
    --cc=linux-nfs@vger.kernel.org \
    --cc=neil@brown.name \
    --cc=okorniev@redhat.com \
    --cc=tom@talpey.com \
    --cc=trond.myklebust@hammerspace.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