linux-kernel - Re: [PATCH v2 0/7] nfsd: CB_SEQUENCE error handling fixes and cleanups

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <Z5o5ZBaYgrmrNseU@poldevia.fieldses.org>
Date: Wed, 29 Jan 2025 09:21:24 -0500
From: "J. Bruce Fields" <bfields@...ldses.org>
To: Jeff Layton <jlayton@...nel.org>
Cc: Chuck Lever <chuck.lever@...cle.com>, Neil Brown <neilb@...e.de>,
	Olga Kornievskaia <okorniev@...hat.com>,
	Dai Ngo <Dai.Ngo@...cle.com>, Tom Talpey <tom@...pey.com>,
	Kinglong Mee <kinglongmee@...il.com>,
	Trond Myklebust <trondmy@...nel.org>,
	Anna Schumaker <anna@...nel.org>, linux-nfs@...r.kernel.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 0/7] nfsd: CB_SEQUENCE error handling fixes and
 cleanups

On Wed, Jan 29, 2025 at 08:39:53AM -0500, Jeff Layton wrote:
> While looking over the CB_SEQUENCE error handling, I discovered that
> callbacks don't hold a reference to a session, and the
> clp->cl_cb_session could easily change between request and response.
> If that happens at an inopportune time, there could be UAFs or weird
> slot/sequence handling problems.

Nobody should place too much faith in my understanding of how any of
this works at this point, but....  My vague memory is that a lot of
things are serialized simply by being run only on the cl_callback_wq.
Modifying clp->cl_cb_session is such a thing.

--b.

> This series changes the nfsd4_session to be RCU-freed, and then adds a
> new method of session refcounting that is compatible with the old.
> nfsd4_callback RPCs will now hold a lightweight reference to the session
> in addition to the slot. Then, all of the callback handling is switched
> to use that session instead of dereferencing clp->cb_cb_session.
> I've also reworked the error handling in nfsd4_cb_sequence_done()
> based on review comments, and lifted the v4.0 handing out of that
> function.
> 
> This passes pynfs, nfstests, and fstests for me, but I'm not sure how
> much any of that stresses the backchannel's error handling.
> 
> These should probably go in via Chuck's tree, but the last patch touches
> some NFS cnd sunrpc client code, so it'd be good to have R-b's or A-b's
> from Trond and/or Anna on that one.
> 
> Signed-off-by: Jeff Layton <jlayton@...nel.org>
> ---
> Changes in v2:
> - make nfsd4_session be RCU-freed
> - change code to keep reference to session over callback RPCs
> - rework error handling in nfsd4_cb_sequence_done()
> - move NFSv4.0 handling out of nfsd4_cb_sequence_done()
> - Link to v1: https://lore.kernel.org/r/20250123-nfsd-6-14-v1-0-c1137a4fa2ae@kernel.org
> 
> ---
> Jeff Layton (7):
>       nfsd: add routines to get/put session references for callbacks
>       nfsd: make clp->cl_cb_session be an RCU managed pointer
>       nfsd: add a cb_ses pointer to nfsd4_callback and use it instead of clp->cb_cb_session
>       nfsd: overhaul CB_SEQUENCE error handling
>       nfsd: remove unneeded forward declaration of nfsd4_mark_cb_fault()
>       nfsd: lift NFSv4.0 handling out of nfsd4_cb_sequence_done()
>       sunrpc: make rpc_restart_call() and rpc_restart_call_prepare() void return
> 
>  fs/nfs/nfs4proc.c           |  12 ++-
>  fs/nfsd/nfs4callback.c      | 212 ++++++++++++++++++++++++++++++++------------
>  fs/nfsd/nfs4state.c         |  43 ++++++++-
>  fs/nfsd/state.h             |   6 +-
>  fs/nfsd/trace.h             |   6 +-
>  include/linux/sunrpc/clnt.h |   4 +-
>  net/sunrpc/clnt.c           |   7 +-
>  7 files changed, 210 insertions(+), 80 deletions(-)
> ---
> base-commit: a05af3c6103b703d1d38d8180b3ebbe0a03c2f07
> change-id: 20250123-nfsd-6-14-b0797e385dc0
> 
> Best regards,
> -- 
> Jeff Layton <jlayton@...nel.org>