From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-alma10-1.taild15c8.ts.net [100.103.45.18]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id EF0322949E0 for ; Sat, 5 Sep 2026 11:03:07 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=100.103.45.18 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1788606189; cv=none; b=mD3Y+tDxVkVoAIMiOK4VfEU6IHTRCxUsTOo0DVihQAKefnTnUikFsWEJXbcjYmO87ZgF3hnOgiiTUSRXkR3ns73bpNWCZFZMj2oChh+TOkN1wWXKnA4IZNiqGczLnWOWEv52c1GZwcfAIEiZZWSPJTK5/wzJLv5OgedjzT59Jyw= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1788606189; c=relaxed/simple; bh=wLRp+w68tv1oMqjtZjA67PqrF7IqHzWsjOJsaGmm+sA=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version:Content-Type; b=Mo+kv27mtbkM615ntE+TNjezzWrUPS3olB72GnCRQtiMtTs59Mgzkbp60lktcxoJZiBlEZvWa0ZtGyl9d6NcTu+rGJegrWNvS0I9Fi1kuT5klH3lHY5Obm+6+f3idL/axqW7AmmByR+ynDYv3Hu+6tDPMNQk5zEFJaqVJd4riNw= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b=V3rsvkOV; arc=none smtp.client-ip=100.103.45.18 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b="V3rsvkOV" Received: by smtp.kernel.org (Postfix) with ESMTPSA id 0CE261F00A3D; Sat, 5 Sep 2026 11:03:06 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=kernel.org; s=k20260515; t=1788606187; bh=Km6wR9TiHhVqM/aa7C2xr+6tNw4nWwJ3G73daYKRaIQ=; h=From:To:Cc:Subject:Date:In-Reply-To:References; b=V3rsvkOVKRBJ0T09KocRq+avkwxpNm6Mv8UbSxNk3tn0ybZlzybbogE0nW76wxc1u 8kNUMNaVp1+TNum8rVP+n+DpIgKnNXWz5vK+rZR9OT0kET+l+fBb6ufUp+jdznl94z xNBvXrCOO6u9ZZs8ZGW5GtSTEEXsNiW3Sy7PhLWd911NnZBqybPHqZsjb5YMAIUKz0 Duz7WsPMQatLlJvmMdXUi0LpzP1lrX6nyebnfKRKlecmNllK8PwdOEr7VmYJbYzxgA psA4uzrDv14NWYnLZ5l9ZvH23hBJNlIWm9p3/QuWBPpSs6s0rexmMOH+nEamRRSfPh fjWeRq/EMO4Mg== From: Sasha Levin To: stable@vger.kernel.org Cc: Jeff Layton , Chris Mason , Chuck Lever , Sasha Levin Subject: [PATCH 6.12.y] nfsd: RCU-protect cl_cb_session to fix use-after-free on session teardown Date: Sat, 5 Sep 2026 07:03:05 -0400 Message-ID: <20260905110305.3663536-1-sashal@kernel.org> X-Mailer: git-send-email 2.53.0 In-Reply-To: <2026090301-reflector-uncrown-1ebb@gregkh> References: <2026090301-reflector-uncrown-1ebb@gregkh> Precedence: bulk X-Mailing-List: stable@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit From: Jeff Layton [ Upstream commit 01c5d5f58a5db9b0ee5afba2e49d3157788687b2 ] After a DESTROY_SESSION the per-session teardown path can free a session while rpciod still holds an inflight callback rpc_task that dereferences clp->cl_cb_session. nfsd4_probe_callback_sync() flushes cl_callback_wq, but once nfsd4_run_cb_work() has called rpc_call_async() the rpc_task lives on rpciod; flushing the workqueue does not wait for it. rpc_shutdown_client() does drain rpciod tasks, but uses a 1-second wait_event_timeout — tasks stuck in rpc_delay() (e.g. 2-second NFS4ERR_DELAY retries) can outlive the drain. destroy path rpciod ------------ ------ unhash_session(ses) nfsd4_probe_callback_sync(clp) flush_workqueue(cl_callback_wq) /* returns; rpc_task still live */ nfsd4_put_session_locked(ses) free_session(ses) -> kfree(ses) nfsd4_cb_sequence_done() reads cb_clp->cl_cb_session /* freed slab */ A second window exists in nfsd4_process_cb_update(). When __nfsd4_find_backchannel() returns NULL because unhash_session() has already removed the destroyed session from cl_sessions, setup_callback_client() takes the v4.1 early return so clp->cl_cb_session = ses never fires and the field retains a pointer to the about-to-be-freed session. Fix both by converting cl_cb_session to an RCU-protected pointer: - Move the cl_cb_session = ses assignment in setup_callback_client() to after rpc_create() succeeds, so it is only published when a working backchannel exists. Clear cl_cb_session on the error return in nfsd4_process_cb_update(). Both stores use rcu_assign_pointer(). - Annotate cl_cb_session with __rcu. All rpciod-side readers use rcu_read_lock()/rcu_dereference() and check for NULL, bailing to the appropriate error or requeue path: encode_cb_sequence4args(), decode_cb_sequence4resok(), nfsd41_cb_get_slot(), nfsd41_cb_release_slot(), nfsd4_cb_prepare(), and nfsd4_cb_sequence_done(). - Switch __free_session() from kfree() to kfree_rcu() so the session slab is not reclaimed until after an RCU grace period, guaranteeing that rpciod readers inside rcu_read_lock() never dereference freed memory. - Pass the session pointer to the nfsd_cb_seq_status and nfsd_cb_free_slot tracepoints instead of having them re-read cl_cb_session. - nfsd4_cb_prepare() calls rpc_exit() when the session is NULL, routing through the done/release path to requeue the callback. Fixes: dcbeaa68dbbd ("nfsd4: allow backchannel recovery") Cc: stable@vger.kernel.org Reported-by: Chris Mason Signed-off-by: Chris Mason Signed-off-by: Jeff Layton Link: https://patch.msgid.link/20260530-nfsd-fixes-v2-2-f27e8eb4d974@kernel.org Signed-off-by: Chuck Lever [ adapted RCU session lifetime handling to the older single-slot callback model and flexible-array session layout ] Signed-off-by: Sasha Levin --- fs/nfsd/nfs4callback.c | 78 +++++++++++++++++++++++++++++++++++------- fs/nfsd/nfs4state.c | 4 +-- fs/nfsd/state.h | 3 +- fs/nfsd/trace.h | 14 ++++---- 4 files changed, 75 insertions(+), 24 deletions(-) diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c index 127626aba7a23..798fcc9c40926 100644 --- a/fs/nfsd/nfs4callback.c +++ b/fs/nfsd/nfs4callback.c @@ -387,12 +387,19 @@ static void encode_cb_sequence4args(struct xdr_stream *xdr, const struct nfsd4_callback *cb, struct nfs4_cb_compound_hdr *hdr) { - struct nfsd4_session *session = cb->cb_clp->cl_cb_session; + struct nfsd4_session *session; __be32 *p; if (hdr->minorversion == 0) return; + rcu_read_lock(); + session = rcu_dereference(cb->cb_clp->cl_cb_session); + if (!session) { + rcu_read_unlock(); + return; + } + encode_nfs_cb_opnum4(xdr, OP_CB_SEQUENCE); encode_sessionid4(xdr, session); @@ -404,6 +411,7 @@ static void encode_cb_sequence4args(struct xdr_stream *xdr, xdr_encode_empty_array(p); /* csa_referring_call_lists */ hdr->nops++; + rcu_read_unlock(); } /* @@ -430,21 +438,32 @@ static void encode_cb_sequence4args(struct xdr_stream *xdr, static int decode_cb_sequence4resok(struct xdr_stream *xdr, struct nfsd4_callback *cb) { - struct nfsd4_session *session = cb->cb_clp->cl_cb_session; + struct nfsd4_session *session; int status = -ESERVERFAULT; __be32 *p; u32 dummy; + rcu_read_lock(); + session = rcu_dereference(cb->cb_clp->cl_cb_session); + if (!session) { + rcu_read_unlock(); + cb->cb_seq_status = -NFS4ERR_BADSESSION; + return -NFS4ERR_BADSESSION; + } + /* * If the server returns different values for sessionID, slotID or * sequence number, the server is looney tunes. */ p = xdr_inline_decode(xdr, NFS4_MAX_SESSIONID_LEN + 4 + 4 + 4 + 4); - if (unlikely(p == NULL)) + if (unlikely(p == NULL)) { + rcu_read_unlock(); goto out_overflow; + } if (memcmp(p, session->se_sessionid.data, NFS4_MAX_SESSIONID_LEN)) { dprintk("NFS: %s Invalid session id\n", __func__); + rcu_read_unlock(); goto out; } p += XDR_QUADLEN(NFS4_MAX_SESSIONID_LEN); @@ -452,18 +471,21 @@ static int decode_cb_sequence4resok(struct xdr_stream *xdr, dummy = be32_to_cpup(p++); if (dummy != session->se_cb_seq_nr) { dprintk("NFS: %s Invalid sequence number\n", __func__); + rcu_read_unlock(); goto out; } dummy = be32_to_cpup(p++); if (dummy != 0) { dprintk("NFS: %s Invalid slotid\n", __func__); + rcu_read_unlock(); goto out; } /* * FIXME: process highest slotid and target highest slotid */ + rcu_read_unlock(); status = 0; out: cb->cb_seq_status = status; @@ -1060,9 +1082,8 @@ static int setup_callback_client(struct nfs4_client *clp, struct nfs4_cb_conn *c } else { if (!conn->cb_xprt || !ses) return -EINVAL; - clp->cl_cb_session = ses; args.bc_xprt = conn->cb_xprt; - args.prognumber = clp->cl_cb_session->se_cb_prog; + args.prognumber = ses->se_cb_prog; args.protocol = conn->cb_xprt->xpt_class->xcl_ident | XPRT_TRANSPORT_BC; args.authflavor = ses->se_cb_sec.flavor; @@ -1080,8 +1101,10 @@ static int setup_callback_client(struct nfs4_client *clp, struct nfs4_cb_conn *c return -ENOMEM; } - if (clp->cl_minorversion != 0) + if (clp->cl_minorversion != 0) { clp->cl_cb_conn.cb_xprt = conn->cb_xprt; + rcu_assign_pointer(clp->cl_cb_session, ses); + } clp->cl_cb_client = client; clp->cl_cb_cred = cred; rcu_read_lock(); @@ -1226,15 +1249,21 @@ static void nfsd4_cb_prepare(struct rpc_task *task, void *calldata) trace_nfsd_cb_rpc_prepare(clp); cb->cb_seq_status = 1; cb->cb_status = 0; - if (minorversion && !nfsd41_cb_get_slot(cb, task)) - return; + if (minorversion) { + if (!rcu_access_pointer(clp->cl_cb_session)) { + rpc_exit(task, -EIO); + return; + } + if (!nfsd41_cb_get_slot(cb, task)) + return; + } rpc_call_start(task); } static bool nfsd4_cb_sequence_done(struct rpc_task *task, struct nfsd4_callback *cb) { struct nfs4_client *clp = cb->cb_clp; - struct nfsd4_session *session = clp->cl_cb_session; + struct nfsd4_session *session; bool ret = true; if (!clp->cl_minorversion) { @@ -1256,8 +1285,15 @@ static bool nfsd4_cb_sequence_done(struct rpc_task *task, struct nfsd4_callback if (!cb->cb_holds_slot) goto need_restart; + rcu_read_lock(); + session = rcu_dereference(clp->cl_cb_session); + if (!session) { + rcu_read_unlock(); + goto need_restart; + } + /* This is the operation status code for CB_SEQUENCE */ - trace_nfsd_cb_seq_status(task, cb); + trace_nfsd_cb_seq_status(task, cb, session); switch (cb->cb_seq_status) { case 0: /* @@ -1285,19 +1321,25 @@ static bool nfsd4_cb_sequence_done(struct rpc_task *task, struct nfsd4_callback case -NFS4ERR_BADSESSION: nfsd4_mark_cb_fault(cb->cb_clp); ret = false; + rcu_read_unlock(); goto need_restart; case -NFS4ERR_DELAY: cb->cb_seq_status = 1; - if (!rpc_restart_call(task)) + if (!rpc_restart_call(task)) { + rcu_read_unlock(); goto out; + } rpc_delay(task, 2 * HZ); + rcu_read_unlock(); return false; case -NFS4ERR_BADSLOT: + rcu_read_unlock(); goto retry_nowait; case -NFS4ERR_SEQ_MISORDERED: if (session->se_cb_seq_nr != 1) { session->se_cb_seq_nr = 1; + rcu_read_unlock(); goto retry_nowait; } break; @@ -1306,7 +1348,8 @@ static bool nfsd4_cb_sequence_done(struct rpc_task *task, struct nfsd4_callback } nfsd41_cb_release_slot(cb); - trace_nfsd_cb_free_slot(task, cb); + trace_nfsd_cb_free_slot(task, cb, session); + rcu_read_unlock(); if (RPC_SIGNALLED(task)) goto need_restart; @@ -1415,7 +1458,15 @@ static struct nfsd4_conn * __nfsd4_find_backchannel(struct nfs4_client *clp) * Note there isn't a lot of locking in this code; instead we depend on * the fact that it is run from clp->cl_callback_wq, which won't run two * work items at once. So, for example, clp->cl_callback_wq handles all - * access of cl_cb_client and all calls to rpc_create or rpc_shutdown_client. + * access of cl_cb_client, and all calls to rpc_create or + * rpc_shutdown_client. + * + * cl_cb_session is written only from cl_callback_wq (via + * rcu_assign_pointer) and read from rpciod under rcu_read_lock (via + * rcu_dereference) by encode_cb_sequence4args(), decode_cb_sequence4resok(), + * and nfsd4_cb_sequence_done(). Sessions are freed with kfree_rcu() so that + * rpciod readers in an RCU read-side critical section never dereference a + * freed session. */ static void nfsd4_process_cb_update(struct nfsd4_callback *cb) { @@ -1467,6 +1518,7 @@ static void nfsd4_process_cb_update(struct nfsd4_callback *cb) nfsd4_mark_cb_down(clp); if (c) svc_xprt_put(c->cn_xprt); + rcu_assign_pointer(clp->cl_cb_session, ses); return; } } diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index 342de8d2c4aa8..7178207afcf2a 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -2152,7 +2152,7 @@ static void nfsd4_del_conns(struct nfsd4_session *s) static void __free_session(struct nfsd4_session *ses) { free_session_slots(ses); - kfree(ses); + kfree_rcu(ses, rcu_head); } static void free_session(struct nfsd4_session *ses) @@ -3195,7 +3195,7 @@ static struct nfs4_client *create_client(struct xdr_netobj name, clear_bit(0, &clp->cl_cb_slot_busy); copy_verf(clp, verf); memcpy(&clp->cl_addr, sa, sizeof(struct sockaddr_storage)); - clp->cl_cb_session = NULL; + RCU_INIT_POINTER(clp->cl_cb_session, NULL); clp->net = net; clp->cl_nfsd_dentry = nfsd_client_mkdir( nn, &clp->cl_nfsdfs, diff --git a/fs/nfsd/state.h b/fs/nfsd/state.h index 5b3a4473ed0ac..75943d7e50e80 100644 --- a/fs/nfsd/state.h +++ b/fs/nfsd/state.h @@ -310,6 +310,7 @@ struct nfsd4_session { struct list_head se_conns; u32 se_cb_prog; u32 se_cb_seq_nr; + struct rcu_head rcu_head; struct nfsd4_slot *se_slots[]; /* forward channel slots */ }; @@ -422,7 +423,7 @@ struct nfs4_client { #define NFSD4_CB_FAULT 3 int cl_cb_state; struct nfsd4_callback cl_cb_null; - struct nfsd4_session *cl_cb_session; + struct nfsd4_session __rcu *cl_cb_session; /* for all client information that callback code might need: */ spinlock_t cl_lock; diff --git a/fs/nfsd/trace.h b/fs/nfsd/trace.h index 3448e444d4100..aed819a3761a1 100644 --- a/fs/nfsd/trace.h +++ b/fs/nfsd/trace.h @@ -1619,9 +1619,10 @@ DEFINE_NFSD_CB_LIFETIME_EVENT(bc_shutdown); TRACE_EVENT(nfsd_cb_seq_status, TP_PROTO( const struct rpc_task *task, - const struct nfsd4_callback *cb + const struct nfsd4_callback *cb, + const struct nfsd4_session *session ), - TP_ARGS(task, cb), + TP_ARGS(task, cb, session), TP_STRUCT__entry( __field(unsigned int, task_id) __field(unsigned int, client_id) @@ -1633,8 +1634,6 @@ TRACE_EVENT(nfsd_cb_seq_status, __field(int, seq_status) ), TP_fast_assign( - const struct nfs4_client *clp = cb->cb_clp; - const struct nfsd4_session *session = clp->cl_cb_session; const struct nfsd4_sessionid *sid = (struct nfsd4_sessionid *)&session->se_sessionid; @@ -1660,9 +1659,10 @@ TRACE_EVENT(nfsd_cb_seq_status, TRACE_EVENT(nfsd_cb_free_slot, TP_PROTO( const struct rpc_task *task, - const struct nfsd4_callback *cb + const struct nfsd4_callback *cb, + const struct nfsd4_session *session ), - TP_ARGS(task, cb), + TP_ARGS(task, cb, session), TP_STRUCT__entry( __field(unsigned int, task_id) __field(unsigned int, client_id) @@ -1673,8 +1673,6 @@ TRACE_EVENT(nfsd_cb_free_slot, __field(u32, slot_seqno) ), TP_fast_assign( - const struct nfs4_client *clp = cb->cb_clp; - const struct nfsd4_session *session = clp->cl_cb_session; const struct nfsd4_sessionid *sid = (struct nfsd4_sessionid *)&session->se_sessionid; -- 2.53.0