nfs: use-after-free in svc_process_common()
Subsystems (4)
A message that is not in the archive (sent off-list or not yet received): [email protected]
PATCH v2 1/4 nfs: use-after-free in svc_process_common()
if node have NFSv41+ mounts inside several net namespaces it can lead to use-after-free in svc_process_common()
svc_process_common()
/* Setup reply header */
rqstp->rq_xprt->xpt_ops->xpo_prep_reply_hdr(rqstp); <<< HEREsvc_process_common() can use already freed rqstp->rq_xprt,
it was assigned in bc_svc_process() where it was taken from serv->sv_bc_xprt. serv is global structure but sv_bc_xprt is assigned per-netnamespace.
To find correct svc_xprt of client-related backchannel bc_svc_process() now calls new .bc_get_xprt callback that executes svc_find_xprt() with proper xprt name.
According to Trond, the whole "let's set up rqstp->rq_xprt for the back channel" is nothing but a giant hack in order to work around the fact that svc_process_common() uses it to find the xpt_ops, and perform a couple of (meaningless for the back channel) tests of xpt_flags.
Trond proposed to pass in the xpt_ops as a new parameter to svc_process_common(), and make those xpt_flags tests check for whether or not rqstp->rq_xprt is actually non-NULL.
It also required to store a pointer to struct net in the struct svc_rqst so that functions called from inside svc_process_common() (nfs4_callback_compound(), svcauth_gss_accept() and some other) can find it. Some other functions was adopted to handle empty rqstp->rq_xprt
--- include/linux/sunrpc/svc.h | 5 ++++- include/linux/sunrpc/xprt.h | 1 + include/trace/events/sunrpc.h | 6 ++++-- net/sunrpc/auth_gss/svcauth_gss.c | 8 +++---- net/sunrpc/svc.c | 36 ++++++++++++++++++++----------- net/sunrpc/svc_xprt.c | 5 +++-- net/sunrpc/xprtrdma/backchannel.c | 5 +++++ net/sunrpc/xprtrdma/transport.c | 1 + net/sunrpc/xprtrdma/xprt_rdma.h | 1 + net/sunrpc/xprtsock.c | 7 ++++++ 10 files changed, 54 insertions(+), 21 deletions(-)
include/linux/sunrpc/svc.h +4 -1
@@ -295,9 +295,12 @@ struct svc_rqst {struct svc_cacherep * rq_cacherep;/* cache info */ struct task_struct *rq_task;/* service thread */ spinlock_t rq_lock;/* per-request lock */ +struct net * rq_bc_net;/* pointer to backchannel's +* net namespace +*/ }; -#define SVC_NET (svc_rqst) (svc_rqst->rq_xprt->xpt_net) +#define SVC_NET (rqst) (rqst->rq_xprt ? rqst->rq_xprt->xpt_net : rqst->rq_bc_net)/* * Rigorous type checking on sockaddr type conversions
include/linux/sunrpc/xprt.h +1 -0
@@ -158,6 +158,7 @@ struct rpc_xprt_ops {int (*bc_setup)(struct rpc_xprt *xprt,unsigned int min_reqs);int (*bc_up)(struct svc_serv *serv,struct net *net); +struct svc_xprt *(*bc_get_xprt)(struct svc_serv *serv,struct net *net);size_t (*bc_maxpayload)(struct rpc_xprt *xprt);void (*bc_free_rqst)(struct rpc_rqst *rqst); void (*bc_destroy)(struct rpc_xprt *xprt,
include/trace/events/sunrpc.h +4 -2
@@ -569,7 +569,8 @@ TRACE_EVENT(svc_process, __field(u32, vers) __field(u32, proc) __string(service, name) -__string (addr, rqst->rq_xprt->xpt_remotebuf) +__string (addr, rqst->rq_xprt ? + rqst->rq_xprt->xpt_remotebuf :"(null)" ) ), TP_fast_assign(
@@ -577,7 +578,8 @@ TRACE_EVENT(svc_process, __entry->vers = rqst->rq_vers; __entry->proc = rqst->rq_proc;__assign_str (service, name); -__assign_str (addr, rqst->rq_xprt->xpt_remotebuf); +__assign_str (addr, rqst->rq_xprt ? + rqst->rq_xprt->xpt_remotebuf :"(null)" ); ), TP_printk("addr=%s xid=0x%08x service=%s vers=%u proc=%u",
net/sunrpc/auth_gss/svcauth_gss.c +4 -4
@@ -1142,7 +1142,7 @@ static int svcauth_gss_legacy_init(struct svc_rqst *rqstp,struct kvec *resv = &rqstp->rq_res.head[0 ];struct rsi *rsip, rsikey;int ret; -struct sunrpc_net *sn =net_generic (rqstp->rq_xprt->xpt_net, sunrpc_net_id); +struct sunrpc_net *sn =net_generic (SVC_NET (rqstp), sunrpc_net_id);memset (&rsikey,0 ,sizeof (rsikey)); ret = gss_read_verf(gc, argv, authp,
@@ -1253,7 +1253,7 @@ static int svcauth_gss_proxy_init(struct svc_rqst *rqstp,uint64_t handle;int status;int ret; -struct net *net = rqstp->rq_xprt->xpt_net; +struct net *net =SVC_NET (rqstp);struct sunrpc_net *sn =net_generic (net, sunrpc_net_id);memset (&ud,0 ,sizeof (ud));
@@ -1444,7 +1444,7 @@ svcauth_gss_accept(struct svc_rqst *rqstp, __be32 *authp)__be32 *rpcstart;__be32 *reject_stat = resv->iov_base + resv->iov_len;int ret; -struct sunrpc_net *sn =net_generic (rqstp->rq_xprt->xpt_net, sunrpc_net_id); +struct sunrpc_net *sn =net_generic (SVC_NET (rqstp), sunrpc_net_id);dprintk ("RPC: svcauth_gss: argv->iov_len = %zd\n" , argv->iov_len);
@@ -1734,7 +1734,7 @@ svcauth_gss_release(struct svc_rqst *rqstp)struct rpc_gss_wire_cred *gc = &gsd->clcred;struct xdr_buf *resbuf = &rqstp->rq_res;int stat = -EINVAL ; -struct sunrpc_net *sn =net_generic (rqstp->rq_xprt->xpt_net, sunrpc_net_id); +struct sunrpc_net *sn =net_generic (SVC_NET (rqstp), sunrpc_net_id);if (gc->gc_proc !=RPC_GSS_PROC_DATA ) goto out;
net/sunrpc/svc.c +24 -12
@@ -1148,7 +1148,8 @@ static __printf(2,3) void svc_printk(struct svc_rqst *rqstp, const char *fmt, ..* Common routine for processing the RPC request. */ static int -svc_process_common (struct svc_rqst *rqstp,struct kvec *argv,struct kvec *resv) +svc_process_common (struct svc_rqst *rqstp,struct kvec *argv, +struct kvec *resv,const struct svc_xprt_ops *xops) {struct svc_program *progp;const struct svc_version *versp =NULL ;/* compiler food */
@@ -1172,7 +1173,7 @@ svc_process_common(struct svc_rqst *rqstp, struct kvec *argv, struct kvec *resv)clear_bit (RQ_DROPME , &rqstp->rq_flags);/* Setup reply header */ - rqstp->rq_xprt->xpt_ops->xpo_prep_reply_hdr (rqstp); + xops->xpo_prep_reply_hdr (rqstp);svc_putu32 (resv, rqstp->rq_xid);
@@ -1244,7 +1245,7 @@ svc_process_common(struct svc_rqst *rqstp, struct kvec *argv, struct kvec *resv)* for lower versions. RPC_PROG_MISMATCH seems to be the closest * fit. */ -if (versp->vs_need_cong_ctrl && +if (versp->vs_need_cong_ctrl && rqstp->rq_xprt && !test_bit (XPT_CONG_CTRL , &rqstp->rq_xprt->xpt_flags)) goto err_bad_vers;
@@ -1336,7 +1337,7 @@ svc_process_common(struct svc_rqst *rqstp, struct kvec *argv, struct kvec *resv)return 0 ; close: -if (test_bit (XPT_TEMP , &rqstp->rq_xprt->xpt_flags)) +if (rqstp->rq_xprt &&test_bit (XPT_TEMP , &rqstp->rq_xprt->xpt_flags))svc_close_xprt (rqstp->rq_xprt);dprintk ("svc: svc_process close\n" );return 0 ;
@@ -1432,7 +1433,8 @@ svc_process(struct svc_rqst *rqstp) }/* Returns 1 for send, 0 for drop */ -if (likely (svc_process_common (rqstp, argv, resv))) +if (likely (svc_process_common (rqstp, argv, resv, + rqstp->rq_xprt->xpt_ops)))return svc_send (rqstp); out_drop:
@@ -1450,19 +1452,25 @@ int bc_svc_process(struct svc_serv *serv, struct rpc_rqst *req, struct svc_rqst *rqstp) { + struct net *net = req->rq_xprt->xprt_net;struct kvec *argv = &rqstp->rq_arg.head[0 ];struct kvec *resv = &rqstp->rq_res.head[0 ];struct rpc_task *task; +struct svc_xprt *s_xprt;int proc_error;int error;dprintk ("svc: %s(%p)\n" , __func__, req); + s_xprt = req->rq_xprt->ops->bc_get_xprt (serv, net); +if (!s_xprt) + goto proc_error; +/* Build the svc_rqst used by the common processing routine */ - rqstp->rq_xprt = serv->sv_bc_xprt; rqstp->rq_xid = req->rq_xid; rqstp->rq_prot = req->rq_xprt->prot; rqstp->rq_server = serv; + rqstp->rq_bc_net = net; rqstp->rq_addrlen =sizeof (req->rq_xprt->addr);memcpy (&rqstp->rq_addr, &req->rq_xprt->addr, rqstp->rq_addrlen);
@@ -1493,14 +1501,12 @@ bc_svc_process(struct svc_serv *serv, struct rpc_rqst *req,svc_getnl (argv);/* CALLDIR */ /* Parse and execute the bc call */ - proc_error =svc_process_common (rqstp, argv, resv); + proc_error =svc_process_common (rqstp, argv, resv, s_xprt->xpt_ops); +svc_xprt_put (s_xprt);atomic_inc (&req->rq_xprt->bc_free_slots); -if (!proc_error) { -/* Processing error: drop the request */ -xprt_free_bc_request (req); -return 0 ; - } +if (!proc_error) + goto proc_error;/* Finally, send the reply synchronously */ memcpy (&req->rq_snd_buf, &rqstp->rq_res,sizeof (req->rq_snd_buf));
@@ -1517,6 +1523,12 @@ bc_svc_process(struct svc_serv *serv, struct rpc_rqst *req, out:dprintk ("svc: %s(), error=%d\n" , __func__, error);return error; + +proc_error: +/* Processing error: drop the request */ +xprt_free_bc_request (req); + error = -EINVAL ; + goto out; }EXPORT_SYMBOL_GPL (bc_svc_process);#endif /* CONFIG_SUNRPC_BACKCHANNEL */
net/sunrpc/svc_xprt.c +3 -2
@@ -468,10 +468,11 @@ static struct svc_xprt *svc_xprt_dequeue(struct svc_pool *pool)*/ void svc_reserve (struct svc_rqst *rqstp,int space) { +struct svc_xprt *xprt = rqstp->rq_xprt; + space += rqstp->rq_res.head[0 ].iov_len; -if (space < rqstp->rq_reserved) { -struct svc_xprt *xprt = rqstp->rq_xprt; +if (xprt && (space < rqstp->rq_reserved)) {atomic_sub ((rqstp->rq_reserved - space), &xprt->xpt_reserved); rqstp->rq_reserved = space;
net/sunrpc/xprtrdma/backchannel.c +5 -0
@@ -133,6 +133,11 @@ int xprt_rdma_bc_up(struct svc_serv *serv, struct net *net)return 0 ; } +struct svc_xprt *xprt_rdma_bc_get_xprt (struct svc_serv *serv,struct net *net) +{ +return svc_find_xprt (serv,"rdma-bc" , net,AF_UNSPEC ,0 ); +} +/** * xprt_rdma_bc_maxpayload - Return maximum backchannel message size * @xprt: transport
net/sunrpc/xprtrdma/transport.c +1 -0
@@ -828,6 +828,7 @@ static const struct rpc_xprt_ops xprt_rdma_procs = {#if defined(CONFIG_SUNRPC_BACKCHANNEL ) .bc_setup = xprt_rdma_bc_setup, .bc_up = xprt_rdma_bc_up, + .bc_get_xprt = xprt_rdma_bc_get_xprt, .bc_maxpayload = xprt_rdma_bc_maxpayload, .bc_free_rqst = xprt_rdma_bc_free_rqst, .bc_destroy = xprt_rdma_bc_destroy,
net/sunrpc/xprtrdma/xprt_rdma.h +1 -0
@@ -662,6 +662,7 @@ void xprt_rdma_cleanup(void);#if defined(CONFIG_SUNRPC_BACKCHANNEL )int xprt_rdma_bc_setup (struct rpc_xprt *,unsigned int );int xprt_rdma_bc_up (struct svc_serv *,struct net *); +struct svc_xprt *xprt_rdma_bc_get_xprt (struct svc_serv *serv,struct net *net);size_t xprt_rdma_bc_maxpayload (struct rpc_xprt *);int rpcrdma_bc_post_recv (struct rpcrdma_xprt *,unsigned int );void rpcrdma_bc_receive_call (struct rpcrdma_xprt *,struct rpcrdma_rep *);
net/sunrpc/xprtsock.c +7 -0
@@ -1411,6 +1411,12 @@ static int xs_tcp_bc_up(struct svc_serv *serv, struct net *net)return 0 ; } +static struct svc_xprt *xs_tcp_bc_get_xprt (struct svc_serv *serv, +struct net *net) +{ +return svc_find_xprt (serv,"tcp-bc" , net,AF_UNSPEC ,0 ); +} +static size_t xs_tcp_bc_maxpayload (struct rpc_xprt *xprt) {return PAGE_SIZE ;
@@ -2668,6 +2674,7 @@ static const struct rpc_xprt_ops xs_tcp_ops = {#ifdef CONFIG_SUNRPC_BACKCHANNEL .bc_setup = xprt_setup_bc, .bc_up = xs_tcp_bc_up, + .bc_get_xprt = xs_tcp_bc_get_xprt, .bc_maxpayload = xs_tcp_bc_maxpayload, .bc_free_rqst = xprt_free_bc_rqst, .bc_destroy = xprt_destroy_bc,
Signature · 1 line
2.17.1
PATCH v2 2/4 nfs: remove sv_bc_enabled using in svc_is_backchannel()
In reply to a message not in the archiveserv->sv_bc_xprt cannot be used as pointer,
serv is global structure, but sv_bc_xprt is assigned per-netnamespace.
The only user of this field is svc_is_backchannel() where it is used not as pointer but as mark of backchannel-compatible servers.
Convert netns-unsafe xprt pointer to simple boolean mark.
--- include/linux/sunrpc/bc_xprt.h | 10 ++++------ include/linux/sunrpc/svc.h | 2 +- net/sunrpc/svcsock.c | 2 +- net/sunrpc/xprtrdma/svc_rdma_transport.c | 2 +- 4 files changed, 7 insertions(+), 9 deletions(-)
include/linux/sunrpc/bc_xprt.h +4 -6
@@ -47,11 +47,9 @@ void xprt_free_bc_rqst(struct rpc_rqst *req);/* * Determine if a shared backchannel is in use */ -static inline int svc_is_backchannel (const struct svc_rqst *rqstp) +static inline bool svc_is_backchannel (const struct svc_rqst *rqstp) { -if (rqstp->rq_server->sv_bc_xprt) -return 1 ; -return 0 ; +return rqstp->rq_server->sv_bc_enabled; }#else /* CONFIG_SUNRPC_BACKCHANNEL */ static inline int xprt_setup_backchannel(struct rpc_xprt *xprt,
@@ -60,9 +58,9 @@ static inline int xprt_setup_backchannel(struct rpc_xprt *xprt,return 0 ; } -static inline int svc_is_backchannel (const struct svc_rqst *rqstp) +static inline bool svc_is_backchannel (const struct svc_rqst *rqstp) { -return 0 ; +return false; }static inline void xprt_free_bc_request (struct rpc_rqst *req)
include/linux/sunrpc/svc.h +1 -1
@@ -109,7 +109,7 @@ struct svc_serv {spinlock_t sv_cb_lock;/* protects the svc_cb_list */ wait_queue_head_t sv_cb_waitq;/* sleep here if there are no * entries in the svc_cb_list */ -struct svc_xprt *sv_bc_xprt;/* callback on fore channel */ +bool sv_bc_enabled;/* server uses backchannel */ #endif /* CONFIG_SUNRPC_BACKCHANNEL */ };
net/sunrpc/svcsock.c +1 -1
@@ -1623,7 +1623,7 @@ static struct svc_xprt *svc_bc_create_socket(struct svc_serv *serv,svc_xprt_init (net, &svc_tcp_bc_class, xprt, serv);set_bit (XPT_CONG_CTRL , &svsk->sk_xprt.xpt_flags); - serv->sv_bc_xprt = xprt; + serv->sv_bc_enabled = true;return xprt; }
net/sunrpc/xprtrdma/svc_rdma_transport.c +1 -1
@@ -136,7 +136,7 @@ static struct svc_xprt *svc_rdma_bc_create(struct svc_serv *serv,svc_xprt_init (net, &svc_rdma_bc_class, xprt, serv);set_bit (XPT_CONG_CTRL , &xprt->xpt_flags); - serv->sv_bc_xprt = xprt; + serv->sv_bc_enabled = true;dprintk ("svcrdma: %s(%p)\n" , __func__, xprt);return xprt;
Signature · 1 line
2.17.1
PATCH v2 3/4 nfs: fix debug message in svc_create_xprt()
In reply to a message not in the archive_svc_create_xprt() returns positive port number
so its non-zero return value is not an error
--- net/sunrpc/svc_xprt.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
net/sunrpc/svc_xprt.c +2 -2
@@ -296,9 +296,9 @@ int svc_create_xprt(struct svc_serv *serv, const char *xprt_name,request_module ("svc%s" , xprt_name); err =_svc_create_xprt (serv, xprt_name, net, family, port, flags); } -if (err) +if (err <0 )dprintk ("svc: transport %s not found, err %d\n" , - xprt_name, err); + xprt_name, -err);return err; }EXPORT_SYMBOL_GPL (svc_create_xprt);
Signature · 1 line
2.17.1
PATCH v2 4/4 nfs: minor typo in nfs4_callback_up_net()
In reply to a message not in the archive--- fs/nfs/callback.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
fs/nfs/callback.c +1 -1
@@ -56,7 +56,7 @@ static int nfs4_callback_up_net(struct svc_serv *serv, struct net *net) nfs_callback_set_tcpport, SVC_SOCK_ANONYMOUS);if (ret >0 ) { nn->nfs_callback_tcpport6 = ret; -dprintk ("NFS: Callback listener port = %u (af %u, net %x\n" , +dprintk ("NFS: Callback listener port = %u (af %u, net %x)\n" , nn->nfs_callback_tcpport6,PF_INET6 , net->ns.inum); }else if (ret != -EAFNOSUPPORT ) goto out_err;
Signature · 1 line
2.17.1