linux/fs/nfs/callback.c
Linus Torvalds 4e4adb2f46 NFS client updates for Linux 4.3
Highlights include:
 
 Stable patches:
 - Fix atomicity of pNFS commit list updates
 - Fix NFSv4 handling of open(O_CREAT|O_EXCL|O_RDONLY)
 - nfs_set_pgio_error sometimes misses errors
 - Fix a thinko in xs_connect()
 - Fix borkage in _same_data_server_addrs_locked()
 - Fix a NULL pointer dereference of migration recovery ops for v4.2 client
 - Don't let the ctime override attribute barriers.
 - Revert "NFSv4: Remove incorrect check in can_open_delegated()"
 - Ensure flexfiles pNFS driver updates the inode after write finishes
 - flexfiles must not pollute the attribute cache with attrbutes from the DS
 - Fix a protocol error in layoutreturn
 - Fix a protocol issue with NFSv4.1 CLOSE stateids
 
 Bugfixes + cleanups
 - pNFS blocks bugfixes from Christoph
 - Various cleanups from Anna
 - More fixes for delegation corner cases
 - Don't fsync twice for O_SYNC/IS_SYNC files
 - Fix pNFS and flexfiles layoutstats bugs
 - pnfs/flexfiles: avoid duplicate tracking of mirror data
 - pnfs: Fix layoutget/layoutreturn/return-on-close serialisation issues.
 - pnfs/flexfiles: error handling retries a layoutget before fallback to MDS
 
 Features:
 - Full support for the OPEN NFS4_CREATE_EXCLUSIVE4_1 mode from Kinglong
 - More RDMA client transport improvements from Chuck
 - Removal of the deprecated ib_reg_phys_mr() and ib_rereg_phys_mr() verbs
   from the SUNRPC, Lustre and core infiniband tree.
 - Optimise away the close-to-open getattr if there is no cached data
 -----BEGIN PGP SIGNATURE-----
 Version: GnuPG v1
 
 iQIcBAABAgAGBQJV7chgAAoJEGcL54qWCgDyqJQP/3kto9VXnXcatC382jF9Pfj5
 F55XeSnviOXH7CyiKA4nSBhnxg/sLuWOTpbkVI/4Y+VyWhLby9h+mtcKURHOlBnj
 d5BFoPwaBVDnUiKlHFQDkRjIyxjj2Sb6/uEb2V/u3v+3znR5AZZ4lzFx4cD85oaz
 mcru7yGiSxaQCIH6lHExcCEKXaDP5YdvS9YFsyQfv2976JSaQHM9ZG04E0v6MzTo
 E5wwC4CLMKmhuX9kmQMj85jzs1ASAKZ3N7b4cApTIo6F8DCDH0vKQphq/nEQC497
 ECjEs5/fpxtNJUpSBu0gT7G4LCiW3PzE7pHa+8bhbaAn9OzxIR5+qWujKsfGYQhO
 Oomp3K9zO6omshAc5w4MkknPpbImjoZjGAj/q/6DbtrDpnD7DzOTirwYY2yX0CA8
 qcL81uJUb8+j4jJj4RTO+lTUBItrM1XTqTSd/3eSMr5DDRVZj+ERZxh17TaxRBZL
 YrbrLHxCHrcbdSbPlovyvY+BwjJUUFJRcOxGQXLmNYR9u92fF59rb53bzVyzcRRO
 wBozzrNRCFL+fPgfNPLEapIb6VtExdM3rl2HYsJGckHj4DPQdnoB3ytIT9iEFZEN
 +/rE14XEZte7kuH3OP4el2UsP/hVsm7A49mtwrkdbd7rxMWD6XfQUp8DAggWUEtI
 1H6T7RC1Y6wsu0X1fnVz
 =knJA
 -----END PGP SIGNATURE-----

Merge tag 'nfs-for-4.3-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

Pull NFS client updates from Trond Myklebust:
 "Highlights include:

  Stable patches:
   - Fix atomicity of pNFS commit list updates
   - Fix NFSv4 handling of open(O_CREAT|O_EXCL|O_RDONLY)
   - nfs_set_pgio_error sometimes misses errors
   - Fix a thinko in xs_connect()
   - Fix borkage in _same_data_server_addrs_locked()
   - Fix a NULL pointer dereference of migration recovery ops for v4.2
     client
   - Don't let the ctime override attribute barriers.
   - Revert "NFSv4: Remove incorrect check in can_open_delegated()"
   - Ensure flexfiles pNFS driver updates the inode after write finishes
   - flexfiles must not pollute the attribute cache with attrbutes from
     the DS
   - Fix a protocol error in layoutreturn
   - Fix a protocol issue with NFSv4.1 CLOSE stateids

  Bugfixes + cleanups
   - pNFS blocks bugfixes from Christoph
   - Various cleanups from Anna
   - More fixes for delegation corner cases
   - Don't fsync twice for O_SYNC/IS_SYNC files
   - Fix pNFS and flexfiles layoutstats bugs
   - pnfs/flexfiles: avoid duplicate tracking of mirror data
   - pnfs: Fix layoutget/layoutreturn/return-on-close serialisation
     issues
   - pnfs/flexfiles: error handling retries a layoutget before fallback
     to MDS

  Features:
   - Full support for the OPEN NFS4_CREATE_EXCLUSIVE4_1 mode from
     Kinglong
   - More RDMA client transport improvements from Chuck
   - Removal of the deprecated ib_reg_phys_mr() and ib_rereg_phys_mr()
     verbs from the SUNRPC, Lustre and core infiniband tree.
   - Optimise away the close-to-open getattr if there is no cached data"

* tag 'nfs-for-4.3-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs: (108 commits)
  NFSv4: Respect the server imposed limit on how many changes we may cache
  NFSv4: Express delegation limit in units of pages
  Revert "NFS: Make close(2) asynchronous when closing NFS O_DIRECT files"
  NFS: Optimise away the close-to-open getattr if there is no cached data
  NFSv4.1/flexfiles: Clean up ff_layout_write_done_cb/ff_layout_commit_done_cb
  NFSv4.1/flexfiles: Mark the layout for return in ff_layout_io_track_ds_error()
  nfs: Remove unneeded checking of the return value from scnprintf
  nfs: Fix truncated client owner id without proto type
  NFSv4.1/flexfiles: Mark layout for return if the mirrors are invalid
  NFSv4.1/flexfiles: RW layouts are valid only if all mirrors are valid
  NFSv4.1/flexfiles: Fix incorrect usage of pnfs_generic_mark_devid_invalid()
  NFSv4.1/flexfiles: Fix freeing of mirrors
  NFSv4.1/pNFS: Don't request a minimal read layout beyond the end of file
  NFSv4.1/pnfs: Handle LAYOUTGET return values correctly
  NFSv4.1/pnfs: Don't ask for a read layout for an empty file.
  NFSv4.1: Fix a protocol issue with CLOSE stateids
  NFSv4.1/flexfiles: Don't mark the entire deviceid as bad for file errors
  SUNRPC: Prevent SYN+SYNACK+RST storms
  SUNRPC: xs_reset_transport must mark the connection as disconnected
  NFSv4.1/pnfs: Ensure layoutreturn reserves space for the opaque payload
  ...
2015-09-07 14:02:24 -07:00

500 lines
12 KiB
C

/*
* linux/fs/nfs/callback.c
*
* Copyright (C) 2004 Trond Myklebust
*
* NFSv4 callback handling
*/
#include <linux/completion.h>
#include <linux/ip.h>
#include <linux/module.h>
#include <linux/sunrpc/svc.h>
#include <linux/sunrpc/svcsock.h>
#include <linux/nfs_fs.h>
#include <linux/errno.h>
#include <linux/mutex.h>
#include <linux/freezer.h>
#include <linux/kthread.h>
#include <linux/sunrpc/svcauth_gss.h>
#include <linux/sunrpc/bc_xprt.h>
#include <net/inet_sock.h>
#include "nfs4_fs.h"
#include "callback.h"
#include "internal.h"
#include "netns.h"
#define NFSDBG_FACILITY NFSDBG_CALLBACK
struct nfs_callback_data {
unsigned int users;
struct svc_serv *serv;
struct svc_rqst *rqst;
struct task_struct *task;
};
static struct nfs_callback_data nfs_callback_info[NFS4_MAX_MINOR_VERSION + 1];
static DEFINE_MUTEX(nfs_callback_mutex);
static struct svc_program nfs4_callback_program;
static int nfs4_callback_up_net(struct svc_serv *serv, struct net *net)
{
int ret;
struct nfs_net *nn = net_generic(net, nfs_net_id);
ret = svc_create_xprt(serv, "tcp", net, PF_INET,
nfs_callback_set_tcpport, SVC_SOCK_ANONYMOUS);
if (ret <= 0)
goto out_err;
nn->nfs_callback_tcpport = ret;
dprintk("NFS: Callback listener port = %u (af %u, net %p)\n",
nn->nfs_callback_tcpport, PF_INET, net);
ret = svc_create_xprt(serv, "tcp", net, PF_INET6,
nfs_callback_set_tcpport, SVC_SOCK_ANONYMOUS);
if (ret > 0) {
nn->nfs_callback_tcpport6 = ret;
dprintk("NFS: Callback listener port = %u (af %u, net %p)\n",
nn->nfs_callback_tcpport6, PF_INET6, net);
} else if (ret != -EAFNOSUPPORT)
goto out_err;
return 0;
out_err:
return (ret) ? ret : -ENOMEM;
}
/*
* This is the NFSv4 callback kernel thread.
*/
static int
nfs4_callback_svc(void *vrqstp)
{
int err;
struct svc_rqst *rqstp = vrqstp;
set_freezable();
while (!kthread_should_stop()) {
/*
* Listen for a request on the socket
*/
err = svc_recv(rqstp, MAX_SCHEDULE_TIMEOUT);
if (err == -EAGAIN || err == -EINTR)
continue;
svc_process(rqstp);
}
return 0;
}
/*
* Prepare to bring up the NFSv4 callback service
*/
static struct svc_rqst *
nfs4_callback_up(struct svc_serv *serv)
{
return svc_prepare_thread(serv, &serv->sv_pools[0], NUMA_NO_NODE);
}
#if defined(CONFIG_NFS_V4_1)
static int nfs41_callback_up_net(struct svc_serv *serv, struct net *net)
{
/*
* Create an svc_sock for the back channel service that shares the
* fore channel connection.
* Returns the input port (0) and sets the svc_serv bc_xprt on success
*/
return svc_create_xprt(serv, "tcp-bc", net, PF_INET, 0,
SVC_SOCK_ANONYMOUS);
}
/*
* The callback service for NFSv4.1 callbacks
*/
static int
nfs41_callback_svc(void *vrqstp)
{
struct svc_rqst *rqstp = vrqstp;
struct svc_serv *serv = rqstp->rq_server;
struct rpc_rqst *req;
int error;
DEFINE_WAIT(wq);
set_freezable();
while (!kthread_should_stop()) {
if (try_to_freeze())
continue;
prepare_to_wait(&serv->sv_cb_waitq, &wq, TASK_INTERRUPTIBLE);
spin_lock_bh(&serv->sv_cb_lock);
if (!list_empty(&serv->sv_cb_list)) {
req = list_first_entry(&serv->sv_cb_list,
struct rpc_rqst, rq_bc_list);
list_del(&req->rq_bc_list);
spin_unlock_bh(&serv->sv_cb_lock);
finish_wait(&serv->sv_cb_waitq, &wq);
dprintk("Invoking bc_svc_process()\n");
error = bc_svc_process(serv, req, rqstp);
dprintk("bc_svc_process() returned w/ error code= %d\n",
error);
} else {
spin_unlock_bh(&serv->sv_cb_lock);
schedule();
finish_wait(&serv->sv_cb_waitq, &wq);
}
flush_signals(current);
}
return 0;
}
/*
* Bring up the NFSv4.1 callback service
*/
static struct svc_rqst *
nfs41_callback_up(struct svc_serv *serv)
{
struct svc_rqst *rqstp;
INIT_LIST_HEAD(&serv->sv_cb_list);
spin_lock_init(&serv->sv_cb_lock);
init_waitqueue_head(&serv->sv_cb_waitq);
rqstp = svc_prepare_thread(serv, &serv->sv_pools[0], NUMA_NO_NODE);
dprintk("--> %s return %d\n", __func__, PTR_ERR_OR_ZERO(rqstp));
return rqstp;
}
static void nfs_minorversion_callback_svc_setup(struct svc_serv *serv,
struct svc_rqst **rqstpp, int (**callback_svc)(void *vrqstp))
{
*rqstpp = nfs41_callback_up(serv);
*callback_svc = nfs41_callback_svc;
}
static inline void nfs_callback_bc_serv(u32 minorversion, struct rpc_xprt *xprt,
struct svc_serv *serv)
{
if (minorversion)
/*
* Save the svc_serv in the transport so that it can
* be referenced when the session backchannel is initialized
*/
xprt->bc_serv = serv;
}
#else
static int nfs41_callback_up_net(struct svc_serv *serv, struct net *net)
{
return 0;
}
static void nfs_minorversion_callback_svc_setup(struct svc_serv *serv,
struct svc_rqst **rqstpp, int (**callback_svc)(void *vrqstp))
{
*rqstpp = ERR_PTR(-ENOTSUPP);
*callback_svc = ERR_PTR(-ENOTSUPP);
}
static inline void nfs_callback_bc_serv(u32 minorversion, struct rpc_xprt *xprt,
struct svc_serv *serv)
{
}
#endif /* CONFIG_NFS_V4_1 */
static int nfs_callback_start_svc(int minorversion, struct rpc_xprt *xprt,
struct svc_serv *serv)
{
struct svc_rqst *rqstp;
int (*callback_svc)(void *vrqstp);
struct nfs_callback_data *cb_info = &nfs_callback_info[minorversion];
int ret;
nfs_callback_bc_serv(minorversion, xprt, serv);
if (cb_info->task)
return 0;
switch (minorversion) {
case 0:
/* v4.0 callback setup */
rqstp = nfs4_callback_up(serv);
callback_svc = nfs4_callback_svc;
break;
default:
nfs_minorversion_callback_svc_setup(serv,
&rqstp, &callback_svc);
}
if (IS_ERR(rqstp))
return PTR_ERR(rqstp);
svc_sock_update_bufs(serv);
cb_info->serv = serv;
cb_info->rqst = rqstp;
cb_info->task = kthread_create(callback_svc, cb_info->rqst,
"nfsv4.%u-svc", minorversion);
if (IS_ERR(cb_info->task)) {
ret = PTR_ERR(cb_info->task);
svc_exit_thread(cb_info->rqst);
cb_info->rqst = NULL;
cb_info->task = NULL;
return ret;
}
rqstp->rq_task = cb_info->task;
wake_up_process(cb_info->task);
dprintk("nfs_callback_up: service started\n");
return 0;
}
static void nfs_callback_down_net(u32 minorversion, struct svc_serv *serv, struct net *net)
{
struct nfs_net *nn = net_generic(net, nfs_net_id);
if (--nn->cb_users[minorversion])
return;
dprintk("NFS: destroy per-net callback data; net=%p\n", net);
svc_shutdown_net(serv, net);
}
static int nfs_callback_up_net(int minorversion, struct svc_serv *serv, struct net *net)
{
struct nfs_net *nn = net_generic(net, nfs_net_id);
int ret;
if (nn->cb_users[minorversion]++)
return 0;
dprintk("NFS: create per-net callback data; net=%p\n", net);
ret = svc_bind(serv, net);
if (ret < 0) {
printk(KERN_WARNING "NFS: bind callback service failed\n");
goto err_bind;
}
switch (minorversion) {
case 0:
ret = nfs4_callback_up_net(serv, net);
break;
case 1:
case 2:
ret = nfs41_callback_up_net(serv, net);
break;
default:
printk(KERN_ERR "NFS: unknown callback version: %d\n",
minorversion);
ret = -EINVAL;
break;
}
if (ret < 0) {
printk(KERN_ERR "NFS: callback service start failed\n");
goto err_socks;
}
return 0;
err_socks:
svc_rpcb_cleanup(serv, net);
err_bind:
dprintk("NFS: Couldn't create callback socket: err = %d; "
"net = %p\n", ret, net);
return ret;
}
static struct svc_serv_ops nfs_cb_sv_ops = {
.svo_enqueue_xprt = svc_xprt_do_enqueue,
};
static struct svc_serv *nfs_callback_create_svc(int minorversion)
{
struct nfs_callback_data *cb_info = &nfs_callback_info[minorversion];
struct svc_serv *serv;
/*
* Check whether we're already up and running.
*/
if (cb_info->task) {
/*
* Note: increase service usage, because later in case of error
* svc_destroy() will be called.
*/
svc_get(cb_info->serv);
return cb_info->serv;
}
/*
* Sanity check: if there's no task,
* we should be the first user ...
*/
if (cb_info->users)
printk(KERN_WARNING "nfs_callback_create_svc: no kthread, %d users??\n",
cb_info->users);
serv = svc_create(&nfs4_callback_program, NFS4_CALLBACK_BUFSIZE, &nfs_cb_sv_ops);
if (!serv) {
printk(KERN_ERR "nfs_callback_create_svc: create service failed\n");
return ERR_PTR(-ENOMEM);
}
/* As there is only one thread we need to over-ride the
* default maximum of 80 connections
*/
serv->sv_maxconn = 1024;
dprintk("nfs_callback_create_svc: service created\n");
return serv;
}
/*
* Bring up the callback thread if it is not already up.
*/
int nfs_callback_up(u32 minorversion, struct rpc_xprt *xprt)
{
struct svc_serv *serv;
struct nfs_callback_data *cb_info = &nfs_callback_info[minorversion];
int ret;
struct net *net = xprt->xprt_net;
mutex_lock(&nfs_callback_mutex);
serv = nfs_callback_create_svc(minorversion);
if (IS_ERR(serv)) {
ret = PTR_ERR(serv);
goto err_create;
}
ret = nfs_callback_up_net(minorversion, serv, net);
if (ret < 0)
goto err_net;
ret = nfs_callback_start_svc(minorversion, xprt, serv);
if (ret < 0)
goto err_start;
cb_info->users++;
/*
* svc_create creates the svc_serv with sv_nrthreads == 1, and then
* svc_prepare_thread increments that. So we need to call svc_destroy
* on both success and failure so that the refcount is 1 when the
* thread exits.
*/
err_net:
svc_destroy(serv);
err_create:
mutex_unlock(&nfs_callback_mutex);
return ret;
err_start:
nfs_callback_down_net(minorversion, serv, net);
dprintk("NFS: Couldn't create server thread; err = %d\n", ret);
goto err_net;
}
/*
* Kill the callback thread if it's no longer being used.
*/
void nfs_callback_down(int minorversion, struct net *net)
{
struct nfs_callback_data *cb_info = &nfs_callback_info[minorversion];
mutex_lock(&nfs_callback_mutex);
nfs_callback_down_net(minorversion, cb_info->serv, net);
cb_info->users--;
if (cb_info->users == 0 && cb_info->task != NULL) {
kthread_stop(cb_info->task);
dprintk("nfs_callback_down: service stopped\n");
svc_exit_thread(cb_info->rqst);
dprintk("nfs_callback_down: service destroyed\n");
cb_info->serv = NULL;
cb_info->rqst = NULL;
cb_info->task = NULL;
}
mutex_unlock(&nfs_callback_mutex);
}
/* Boolean check of RPC_AUTH_GSS principal */
int
check_gss_callback_principal(struct nfs_client *clp, struct svc_rqst *rqstp)
{
char *p = rqstp->rq_cred.cr_principal;
if (rqstp->rq_authop->flavour != RPC_AUTH_GSS)
return 1;
/* No RPC_AUTH_GSS on NFSv4.1 back channel yet */
if (clp->cl_minorversion != 0)
return 0;
/*
* It might just be a normal user principal, in which case
* userspace won't bother to tell us the name at all.
*/
if (p == NULL)
return 0;
/*
* Did we get the acceptor from userland during the SETCLIENID
* negotiation?
*/
if (clp->cl_acceptor)
return !strcmp(p, clp->cl_acceptor);
/*
* Otherwise try to verify it using the cl_hostname. Note that this
* doesn't work if a non-canonical hostname was used in the devname.
*/
/* Expect a GSS_C_NT_HOSTBASED_NAME like "nfs@serverhostname" */
if (memcmp(p, "nfs@", 4) != 0)
return 0;
p += 4;
if (strcmp(p, clp->cl_hostname) != 0)
return 0;
return 1;
}
/*
* pg_authenticate method for nfsv4 callback threads.
*
* The authflavor has been negotiated, so an incorrect flavor is a server
* bug. Deny packets with incorrect authflavor.
*
* All other checking done after NFS decoding where the nfs_client can be
* found in nfs4_callback_compound
*/
static int nfs_callback_authenticate(struct svc_rqst *rqstp)
{
switch (rqstp->rq_authop->flavour) {
case RPC_AUTH_NULL:
if (rqstp->rq_proc != CB_NULL)
return SVC_DENIED;
break;
case RPC_AUTH_GSS:
/* No RPC_AUTH_GSS support yet in NFSv4.1 */
if (svc_is_backchannel(rqstp))
return SVC_DENIED;
}
return SVC_OK;
}
/*
* Define NFS4 callback program
*/
static struct svc_version *nfs4_callback_version[] = {
[1] = &nfs4_callback_version1,
[4] = &nfs4_callback_version4,
};
static struct svc_stat nfs4_callback_stats;
static struct svc_program nfs4_callback_program = {
.pg_prog = NFS4_CALLBACK, /* RPC service number */
.pg_nvers = ARRAY_SIZE(nfs4_callback_version), /* Number of entries */
.pg_vers = nfs4_callback_version, /* version table */
.pg_name = "NFSv4 callback", /* service name */
.pg_class = "nfs", /* authentication class */
.pg_stats = &nfs4_callback_stats,
.pg_authenticate = nfs_callback_authenticate,
};