From: Jeff Layton <jlayton@kernel.org>
To: Dominik Maier <dmnk@google.com>, Chuck Lever <cel@kernel.org>
Cc: NeilBrown <neil@brown.name>,
Olga Kornievskaia <okorniev@redhat.com>,
Dai Ngo <Dai.Ngo@oracle.com>, Tom Talpey <tom@talpey.com>,
linux-nfs@vger.kernel.org, dmnk+artist@google.com
Subject: Re: [PATCH] nfsd: only set NFSD_MAY_BYPASS_GSS for RPC_AUTH_UNIX in nlm_fopen()
Date: Thu, 19 Nov 2026 13:45:15 -0800 [thread overview]
Message-ID: <a36e3bd7fdca210e482f14eb0f0daedd95c074b7.camel@kernel.org> (raw)
In-Reply-To: <20261119192336.2697518-1-dmnk@google.com>
On Thu, 2026-11-19 at 20:23 +0100, Dominik Maier wrote:
> As documented in nlm_fopen(), NFSD_MAY_BYPASS_GSS is intended to let NLM
> clients using AUTH_SYS (RPC_AUTH_UNIX) lock files on Kerberos NFS
> exports, whereas NLM clients using AUTH_NULL (RPC_AUTH_NULL) are only
> meant to bypass authentication when NFSEXP_NOAUTHNLM (no_auth_nlm) is
> enabled via NFSD_MAY_NLM.
>
> However, nlm_fopen() has historically passed NFSD_MAY_BYPASS_GSS
> unconditionally, and check_security_flavor() accepts both RPC_AUTH_UNIX
> and RPC_AUTH_NULL whenever may_bypass_gss is true on an export with a
> pseudoflavor >= RPC_AUTH_DES. Since commit bb4f07f2409c ("nfsd: Fix
> NFSD_MAY_BYPASS_GSS and NFSD_MAY_BYPASS_GSS_ON_ROOT") restricted
> may_bypass_gss to exports with pseudoflavor >= RPC_AUTH_DES, default
> auth_nlm rejects RPC_AUTH_NULL NLM requests on sec=sys exports with
> nfserr_wrongsec while still accepting RPC_AUTH_NULL NLM requests on
> sec=krb5* exports even when no_auth_nlm is not set.
>
> Only set NFSD_MAY_BYPASS_GSS in nlm_fopen() when rq_cred.cr_flavor is
> RPC_AUTH_UNIX so auth_nlm and no_auth_nlm behave consistently across
> sec=sys and GSS exports, requiring no_auth_nlm or an explicit sec=none
> export flavor for RPC_AUTH_NULL NLM requests.
>
> This issue was discovered by an AI-assisted code review tool (ARTIST),
> validated in QEMU against an unmodified v7.4-rc3 kernel without kernel
> modifications or error injection, build-tested with KCFLAGS=-Werror, and
> checked with checkpatch.pl --strict.
>
Sashiko suggests:
Fixes: 4cc9b9f2bf4d ("nfsd: refine and rename NFSD_MAY_LOCK")
> Signed-off-by: Dominik Maier <dmnk@google.com>
> ---
> fs/nfsd/lockd.c | 8 +++++---
> 1 file changed, 5 insertions(+), 3 deletions(-)
>
> diff --git a/fs/nfsd/lockd.c b/fs/nfsd/lockd.c
> index f24e45d..2683939 100644
> --- a/fs/nfsd/lockd.c
> +++ b/fs/nfsd/lockd.c
> @@ -41,8 +41,8 @@ static int nlm_fopen(struct svc_rqst *rqstp, struct nfs_fh *f,
> fh.fh_export = NULL;
>
> /*
> - * Allow BYPASS_GSS as some client implementations use AUTH_SYS
> - * for NLM even when GSS is used for NFS.
> + * Allow BYPASS_GSS for AUTH_SYS as some client implementations use
> + * AUTH_SYS for NLM even when GSS is used for NFS.
> * Allow OWNER_OVERRIDE as permission might have been changed
> * after the file was opened.
> * Pass MAY_NLM so that authentication can be completely bypassed
> @@ -50,7 +50,9 @@ static int nlm_fopen(struct svc_rqst *rqstp, struct nfs_fh *f,
> * for NLM requests.
> */
> access = (flags == O_WRONLY) ? NFSD_MAY_WRITE : NFSD_MAY_READ;
> - access |= NFSD_MAY_NLM | NFSD_MAY_OWNER_OVERRIDE | NFSD_MAY_BYPASS_GSS;
> + access |= NFSD_MAY_NLM | NFSD_MAY_OWNER_OVERRIDE;
> + if (rqstp->rq_cred.cr_flavor == RPC_AUTH_UNIX)
> + access |= NFSD_MAY_BYPASS_GSS;
> nfserr = nfsd_open(rqstp, &fh, S_IFREG, access, filp);
> fh_put(&fh);
>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
prev parent reply other threads:[~2026-11-19 21:45 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-19 19:23 [PATCH] nfsd: only set NFSD_MAY_BYPASS_GSS for RPC_AUTH_UNIX in nlm_fopen() Dominik Maier
2026-11-19 21:45 ` Jeff Layton [this message]
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=a36e3bd7fdca210e482f14eb0f0daedd95c074b7.camel@kernel.org \
--to=jlayton@kernel.org \
--cc=Dai.Ngo@oracle.com \
--cc=cel@kernel.org \
--cc=dmnk+artist@google.com \
--cc=dmnk@google.com \
--cc=linux-nfs@vger.kernel.org \
--cc=neil@brown.name \
--cc=okorniev@redhat.com \
--cc=tom@talpey.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