linux/fs/nfs/nfs4proc.c

3739 lines
99 KiB
C
Raw Normal View History

/*
* fs/nfs/nfs4proc.c
*
* Client-side procedure declarations for NFSv4.
*
* Copyright (c) 2002 The Regents of the University of Michigan.
* All rights reserved.
*
* Kendrick Smith <kmsmith@umich.edu>
* Andy Adamson <andros@umich.edu>
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
*
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. Neither the name of the University nor the names of its
* contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED
* WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
* DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
* BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
* LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
* NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
* SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#include <linux/mm.h>
#include <linux/utsname.h>
#include <linux/delay.h>
#include <linux/errno.h>
#include <linux/string.h>
#include <linux/sunrpc/clnt.h>
#include <linux/nfs.h>
#include <linux/nfs4.h>
#include <linux/nfs_fs.h>
#include <linux/nfs_page.h>
#include <linux/smp_lock.h>
#include <linux/namei.h>
#include <linux/mount.h>
#include "nfs4_fs.h"
#include "delegation.h"
#include "internal.h"
#include "iostat.h"
#define NFSDBG_FACILITY NFSDBG_PROC
#define NFS4_POLL_RETRY_MIN (HZ/10)
#define NFS4_POLL_RETRY_MAX (15*HZ)
struct nfs4_opendata;
static int _nfs4_proc_open(struct nfs4_opendata *data);
static int nfs4_do_fsinfo(struct nfs_server *, struct nfs_fh *, struct nfs_fsinfo *);
static int nfs4_async_handle_error(struct rpc_task *, const struct nfs_server *);
static int nfs4_handle_exception(const struct nfs_server *server, int errorcode, struct nfs4_exception *exception);
static int nfs4_wait_clnt_recover(struct rpc_clnt *clnt, struct nfs_client *clp);
static int _nfs4_proc_lookup(struct inode *dir, const struct qstr *name, struct nfs_fh *fhandle, struct nfs_fattr *fattr);
static int _nfs4_proc_getattr(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fattr *fattr);
/* Prevent leaks of NFSv4 errors into userland */
int nfs4_map_errors(int err)
{
if (err < -1000) {
dprintk("%s could not handle NFSv4 error %d\n",
__func__, -err);
return -EIO;
}
return err;
}
/*
* This is our standard bitmap for GETATTR requests.
*/
const u32 nfs4_fattr_bitmap[2] = {
FATTR4_WORD0_TYPE
| FATTR4_WORD0_CHANGE
| FATTR4_WORD0_SIZE
| FATTR4_WORD0_FSID
| FATTR4_WORD0_FILEID,
FATTR4_WORD1_MODE
| FATTR4_WORD1_NUMLINKS
| FATTR4_WORD1_OWNER
| FATTR4_WORD1_OWNER_GROUP
| FATTR4_WORD1_RAWDEV
| FATTR4_WORD1_SPACE_USED
| FATTR4_WORD1_TIME_ACCESS
| FATTR4_WORD1_TIME_METADATA
| FATTR4_WORD1_TIME_MODIFY
};
const u32 nfs4_statfs_bitmap[2] = {
FATTR4_WORD0_FILES_AVAIL
| FATTR4_WORD0_FILES_FREE
| FATTR4_WORD0_FILES_TOTAL,
FATTR4_WORD1_SPACE_AVAIL
| FATTR4_WORD1_SPACE_FREE
| FATTR4_WORD1_SPACE_TOTAL
};
const u32 nfs4_pathconf_bitmap[2] = {
FATTR4_WORD0_MAXLINK
| FATTR4_WORD0_MAXNAME,
0
};
const u32 nfs4_fsinfo_bitmap[2] = { FATTR4_WORD0_MAXFILESIZE
| FATTR4_WORD0_MAXREAD
| FATTR4_WORD0_MAXWRITE
| FATTR4_WORD0_LEASE_TIME,
0
};
const u32 nfs4_fs_locations_bitmap[2] = {
FATTR4_WORD0_TYPE
| FATTR4_WORD0_CHANGE
| FATTR4_WORD0_SIZE
| FATTR4_WORD0_FSID
| FATTR4_WORD0_FILEID
| FATTR4_WORD0_FS_LOCATIONS,
FATTR4_WORD1_MODE
| FATTR4_WORD1_NUMLINKS
| FATTR4_WORD1_OWNER
| FATTR4_WORD1_OWNER_GROUP
| FATTR4_WORD1_RAWDEV
| FATTR4_WORD1_SPACE_USED
| FATTR4_WORD1_TIME_ACCESS
| FATTR4_WORD1_TIME_METADATA
| FATTR4_WORD1_TIME_MODIFY
| FATTR4_WORD1_MOUNTED_ON_FILEID
};
static void nfs4_setup_readdir(u64 cookie, __be32 *verifier, struct dentry *dentry,
struct nfs4_readdir_arg *readdir)
{
__be32 *start, *p;
BUG_ON(readdir->count < 80);
if (cookie > 2) {
readdir->cookie = cookie;
memcpy(&readdir->verifier, verifier, sizeof(readdir->verifier));
return;
}
readdir->cookie = 0;
memset(&readdir->verifier, 0, sizeof(readdir->verifier));
if (cookie == 2)
return;
/*
* NFSv4 servers do not return entries for '.' and '..'
* Therefore, we fake these entries here. We let '.'
* have cookie 0 and '..' have cookie 1. Note that
* when talking to the server, we always send cookie 0
* instead of 1 or 2.
*/
start = p = kmap_atomic(*readdir->pages, KM_USER0);
if (cookie == 0) {
*p++ = xdr_one; /* next */
*p++ = xdr_zero; /* cookie, first word */
*p++ = xdr_one; /* cookie, second word */
*p++ = xdr_one; /* entry len */
memcpy(p, ".\0\0\0", 4); /* entry */
p++;
*p++ = xdr_one; /* bitmap length */
*p++ = htonl(FATTR4_WORD0_FILEID); /* bitmap */
*p++ = htonl(8); /* attribute buffer length */
p = xdr_encode_hyper(p, NFS_FILEID(dentry->d_inode));
}
*p++ = xdr_one; /* next */
*p++ = xdr_zero; /* cookie, first word */
*p++ = xdr_two; /* cookie, second word */
*p++ = xdr_two; /* entry len */
memcpy(p, "..\0\0", 4); /* entry */
p++;
*p++ = xdr_one; /* bitmap length */
*p++ = htonl(FATTR4_WORD0_FILEID); /* bitmap */
*p++ = htonl(8); /* attribute buffer length */
p = xdr_encode_hyper(p, NFS_FILEID(dentry->d_parent->d_inode));
readdir->pgbase = (char *)p - (char *)start;
readdir->count -= readdir->pgbase;
kunmap_atomic(start, KM_USER0);
}
static void renew_lease(const struct nfs_server *server, unsigned long timestamp)
{
struct nfs_client *clp = server->nfs_client;
spin_lock(&clp->cl_lock);
if (time_before(clp->cl_last_renewal,timestamp))
clp->cl_last_renewal = timestamp;
spin_unlock(&clp->cl_lock);
}
static void update_changeattr(struct inode *dir, struct nfs4_change_info *cinfo)
{
struct nfs_inode *nfsi = NFS_I(dir);
spin_lock(&dir->i_lock);
nfsi->cache_validity |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE|NFS_INO_INVALID_DATA;
if (!cinfo->atomic || cinfo->before != nfsi->change_attr)
nfs_force_lookup_revalidate(dir);
nfsi->change_attr = cinfo->after;
spin_unlock(&dir->i_lock);
}
struct nfs4_opendata {
struct kref kref;
struct nfs_openargs o_arg;
struct nfs_openres o_res;
struct nfs_open_confirmargs c_arg;
struct nfs_open_confirmres c_res;
struct nfs_fattr f_attr;
struct nfs_fattr dir_attr;
struct path path;
struct dentry *dir;
struct nfs4_state_owner *owner;
struct nfs4_state *state;
struct iattr attrs;
unsigned long timestamp;
unsigned int rpc_done : 1;
int rpc_status;
int cancelled;
};
static void nfs4_init_opendata_res(struct nfs4_opendata *p)
{
p->o_res.f_attr = &p->f_attr;
p->o_res.dir_attr = &p->dir_attr;
p->o_res.seqid = p->o_arg.seqid;
p->c_res.seqid = p->c_arg.seqid;
p->o_res.server = p->o_arg.server;
nfs_fattr_init(&p->f_attr);
nfs_fattr_init(&p->dir_attr);
}
static struct nfs4_opendata *nfs4_opendata_alloc(struct path *path,
struct nfs4_state_owner *sp, int flags,
const struct iattr *attrs)
{
struct dentry *parent = dget_parent(path->dentry);
struct inode *dir = parent->d_inode;
struct nfs_server *server = NFS_SERVER(dir);
struct nfs4_opendata *p;
p = kzalloc(sizeof(*p), GFP_KERNEL);
if (p == NULL)
goto err;
p->o_arg.seqid = nfs_alloc_seqid(&sp->so_seqid);
if (p->o_arg.seqid == NULL)
goto err_free;
p->path.mnt = mntget(path->mnt);
p->path.dentry = dget(path->dentry);
p->dir = parent;
p->owner = sp;
atomic_inc(&sp->so_count);
p->o_arg.fh = NFS_FH(dir);
p->o_arg.open_flags = flags,
p->o_arg.clientid = server->nfs_client->cl_clientid;
p->o_arg.id = sp->so_owner_id.id;
p->o_arg.name = &p->path.dentry->d_name;
p->o_arg.server = server;
p->o_arg.bitmask = server->attr_bitmask;
p->o_arg.claim = NFS4_OPEN_CLAIM_NULL;
if (flags & O_EXCL) {
u32 *s = (u32 *) p->o_arg.u.verifier.data;
s[0] = jiffies;
s[1] = current->pid;
} else if (flags & O_CREAT) {
p->o_arg.u.attrs = &p->attrs;
memcpy(&p->attrs, attrs, sizeof(p->attrs));
}
p->c_arg.fh = &p->o_res.fh;
p->c_arg.stateid = &p->o_res.stateid;
p->c_arg.seqid = p->o_arg.seqid;
nfs4_init_opendata_res(p);
kref_init(&p->kref);
return p;
err_free:
kfree(p);
err:
dput(parent);
return NULL;
}
static void nfs4_opendata_free(struct kref *kref)
{
struct nfs4_opendata *p = container_of(kref,
struct nfs4_opendata, kref);
nfs_free_seqid(p->o_arg.seqid);
if (p->state != NULL)
nfs4_put_open_state(p->state);
nfs4_put_state_owner(p->owner);
dput(p->dir);
path_put(&p->path);
kfree(p);
}
static void nfs4_opendata_put(struct nfs4_opendata *p)
{
if (p != NULL)
kref_put(&p->kref, nfs4_opendata_free);
}
static int nfs4_wait_for_completion_rpc_task(struct rpc_task *task)
{
int ret;
ret = rpc_wait_for_completion_task(task);
return ret;
}
static int can_open_cached(struct nfs4_state *state, int mode)
{
int ret = 0;
switch (mode & (FMODE_READ|FMODE_WRITE|O_EXCL)) {
case FMODE_READ:
ret |= test_bit(NFS_O_RDONLY_STATE, &state->flags) != 0;
break;
case FMODE_WRITE:
ret |= test_bit(NFS_O_WRONLY_STATE, &state->flags) != 0;
break;
case FMODE_READ|FMODE_WRITE:
ret |= test_bit(NFS_O_RDWR_STATE, &state->flags) != 0;
}
return ret;
}
static int can_open_delegated(struct nfs_delegation *delegation, mode_t open_flags)
{
if ((delegation->type & open_flags) != open_flags)
return 0;
if (delegation->flags & NFS_DELEGATION_NEED_RECLAIM)
return 0;
return 1;
}
static void update_open_stateflags(struct nfs4_state *state, mode_t open_flags)
{
switch (open_flags) {
case FMODE_WRITE:
state->n_wronly++;
break;
case FMODE_READ:
state->n_rdonly++;
break;
case FMODE_READ|FMODE_WRITE:
state->n_rdwr++;
}
nfs4_state_set_mode_locked(state, state->state | open_flags);
}
static void nfs_set_open_stateid_locked(struct nfs4_state *state, nfs4_stateid *stateid, int open_flags)
{
if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0)
memcpy(state->stateid.data, stateid->data, sizeof(state->stateid.data));
memcpy(state->open_stateid.data, stateid->data, sizeof(state->open_stateid.data));
switch (open_flags) {
case FMODE_READ:
set_bit(NFS_O_RDONLY_STATE, &state->flags);
break;
case FMODE_WRITE:
set_bit(NFS_O_WRONLY_STATE, &state->flags);
break;
case FMODE_READ|FMODE_WRITE:
set_bit(NFS_O_RDWR_STATE, &state->flags);
}
}
static void nfs_set_open_stateid(struct nfs4_state *state, nfs4_stateid *stateid, int open_flags)
{
write_seqlock(&state->seqlock);
nfs_set_open_stateid_locked(state, stateid, open_flags);
write_sequnlock(&state->seqlock);
}
static void update_open_stateid(struct nfs4_state *state, nfs4_stateid *open_stateid, nfs4_stateid *deleg_stateid, int open_flags)
{
open_flags &= (FMODE_READ|FMODE_WRITE);
/*
* Protect the call to nfs4_state_set_mode_locked and
* serialise the stateid update
*/
write_seqlock(&state->seqlock);
if (deleg_stateid != NULL) {
memcpy(state->stateid.data, deleg_stateid->data, sizeof(state->stateid.data));
set_bit(NFS_DELEGATED_STATE, &state->flags);
}
if (open_stateid != NULL)
nfs_set_open_stateid_locked(state, open_stateid, open_flags);
write_sequnlock(&state->seqlock);
spin_lock(&state->owner->so_lock);
update_open_stateflags(state, open_flags);
spin_unlock(&state->owner->so_lock);
}
static void nfs4_return_incompatible_delegation(struct inode *inode, mode_t open_flags)
{
struct nfs_delegation *delegation;
rcu_read_lock();
delegation = rcu_dereference(NFS_I(inode)->delegation);
if (delegation == NULL || (delegation->type & open_flags) == open_flags) {
rcu_read_unlock();
return;
}
rcu_read_unlock();
nfs_inode_return_delegation(inode);
}
static struct nfs4_state *nfs4_try_open_cached(struct nfs4_opendata *opendata)
{
struct nfs4_state *state = opendata->state;
struct nfs_inode *nfsi = NFS_I(state->inode);
struct nfs_delegation *delegation;
int open_mode = opendata->o_arg.open_flags & (FMODE_READ|FMODE_WRITE|O_EXCL);
nfs4_stateid stateid;
int ret = -EAGAIN;
rcu_read_lock();
delegation = rcu_dereference(nfsi->delegation);
for (;;) {
if (can_open_cached(state, open_mode)) {
spin_lock(&state->owner->so_lock);
if (can_open_cached(state, open_mode)) {
update_open_stateflags(state, open_mode);
spin_unlock(&state->owner->so_lock);
rcu_read_unlock();
goto out_return_state;
}
spin_unlock(&state->owner->so_lock);
}
if (delegation == NULL)
break;
if (!can_open_delegated(delegation, open_mode))
break;
/* Save the delegation */
memcpy(stateid.data, delegation->stateid.data, sizeof(stateid.data));
rcu_read_unlock();
lock_kernel();
ret = nfs_may_open(state->inode, state->owner->so_cred, open_mode);
unlock_kernel();
if (ret != 0)
goto out;
ret = -EAGAIN;
rcu_read_lock();
delegation = rcu_dereference(nfsi->delegation);
/* If no delegation, try a cached open */
if (delegation == NULL)
continue;
/* Is the delegation still valid? */
if (memcmp(stateid.data, delegation->stateid.data, sizeof(stateid.data)) != 0)
continue;
rcu_read_unlock();
update_open_stateid(state, NULL, &stateid, open_mode);
goto out_return_state;
}
rcu_read_unlock();
out:
return ERR_PTR(ret);
out_return_state:
atomic_inc(&state->count);
return state;
}
static struct nfs4_state *nfs4_opendata_to_nfs4_state(struct nfs4_opendata *data)
{
struct inode *inode;
struct nfs4_state *state = NULL;
struct nfs_delegation *delegation;
nfs4_stateid *deleg_stateid = NULL;
int ret;
if (!data->rpc_done) {
state = nfs4_try_open_cached(data);
goto out;
}
ret = -EAGAIN;
if (!(data->f_attr.valid & NFS_ATTR_FATTR))
goto err;
inode = nfs_fhget(data->dir->d_sb, &data->o_res.fh, &data->f_attr);
ret = PTR_ERR(inode);
if (IS_ERR(inode))
goto err;
ret = -ENOMEM;
state = nfs4_get_open_state(inode, data->owner);
if (state == NULL)
goto err_put_inode;
if (data->o_res.delegation_type != 0) {
int delegation_flags = 0;
rcu_read_lock();
delegation = rcu_dereference(NFS_I(inode)->delegation);
if (delegation)
delegation_flags = delegation->flags;
rcu_read_unlock();
if (!(delegation_flags & NFS_DELEGATION_NEED_RECLAIM))
nfs_inode_set_delegation(state->inode,
data->owner->so_cred,
&data->o_res);
else
nfs_inode_reclaim_delegation(state->inode,
data->owner->so_cred,
&data->o_res);
}
rcu_read_lock();
delegation = rcu_dereference(NFS_I(inode)->delegation);
if (delegation != NULL)
deleg_stateid = &delegation->stateid;
update_open_stateid(state, &data->o_res.stateid, deleg_stateid, data->o_arg.open_flags);
rcu_read_unlock();
iput(inode);
out:
return state;
err_put_inode:
iput(inode);
err:
return ERR_PTR(ret);
}
static struct nfs_open_context *nfs4_state_find_open_context(struct nfs4_state *state)
{
struct nfs_inode *nfsi = NFS_I(state->inode);
struct nfs_open_context *ctx;
spin_lock(&state->inode->i_lock);
list_for_each_entry(ctx, &nfsi->open_files, list) {
if (ctx->state != state)
continue;
get_nfs_open_context(ctx);
spin_unlock(&state->inode->i_lock);
return ctx;
}
spin_unlock(&state->inode->i_lock);
return ERR_PTR(-ENOENT);
}
static struct nfs4_opendata *nfs4_open_recoverdata_alloc(struct nfs_open_context *ctx, struct nfs4_state *state)
{
struct nfs4_opendata *opendata;
opendata = nfs4_opendata_alloc(&ctx->path, state->owner, 0, NULL);
if (opendata == NULL)
return ERR_PTR(-ENOMEM);
opendata->state = state;
atomic_inc(&state->count);
return opendata;
}
static int nfs4_open_recover_helper(struct nfs4_opendata *opendata, mode_t openflags, struct nfs4_state **res)
{
struct nfs4_state *newstate;
int ret;
opendata->o_arg.open_flags = openflags;
memset(&opendata->o_res, 0, sizeof(opendata->o_res));
memset(&opendata->c_res, 0, sizeof(opendata->c_res));
nfs4_init_opendata_res(opendata);
ret = _nfs4_proc_open(opendata);
if (ret != 0)
return ret;
newstate = nfs4_opendata_to_nfs4_state(opendata);
if (IS_ERR(newstate))
return PTR_ERR(newstate);
nfs4_close_state(&opendata->path, newstate, openflags);
*res = newstate;
return 0;
}
static int nfs4_open_recover(struct nfs4_opendata *opendata, struct nfs4_state *state)
{
struct nfs4_state *newstate;
int ret;
/* memory barrier prior to reading state->n_* */
clear_bit(NFS_DELEGATED_STATE, &state->flags);
smp_rmb();
if (state->n_rdwr != 0) {
ret = nfs4_open_recover_helper(opendata, FMODE_READ|FMODE_WRITE, &newstate);
if (ret != 0)
return ret;
if (newstate != state)
return -ESTALE;
}
if (state->n_wronly != 0) {
ret = nfs4_open_recover_helper(opendata, FMODE_WRITE, &newstate);
if (ret != 0)
return ret;
if (newstate != state)
return -ESTALE;
}
if (state->n_rdonly != 0) {
ret = nfs4_open_recover_helper(opendata, FMODE_READ, &newstate);
if (ret != 0)
return ret;
if (newstate != state)
return -ESTALE;
}
/*
* We may have performed cached opens for all three recoveries.
* Check if we need to update the current stateid.
*/
if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0 &&
memcmp(state->stateid.data, state->open_stateid.data, sizeof(state->stateid.data)) != 0) {
write_seqlock(&state->seqlock);
if (test_bit(NFS_DELEGATED_STATE, &state->flags) == 0)
memcpy(state->stateid.data, state->open_stateid.data, sizeof(state->stateid.data));
write_sequnlock(&state->seqlock);
}
return 0;
}
/*
* OPEN_RECLAIM:
* reclaim state on the server after a reboot.
*/
static int _nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state *state)
{
struct nfs_delegation *delegation;
struct nfs4_opendata *opendata;
int delegation_type = 0;
int status;
opendata = nfs4_open_recoverdata_alloc(ctx, state);
if (IS_ERR(opendata))
return PTR_ERR(opendata);
opendata->o_arg.claim = NFS4_OPEN_CLAIM_PREVIOUS;
opendata->o_arg.fh = NFS_FH(state->inode);
rcu_read_lock();
delegation = rcu_dereference(NFS_I(state->inode)->delegation);
if (delegation != NULL && (delegation->flags & NFS_DELEGATION_NEED_RECLAIM) != 0)
NFSv4: Fix a typo in _nfs4_do_open_reclaim This should fix the following Oops reported by Jeff Garzik: kernel BUG at fs/nfs/nfs4xdr.c:1040! invalid opcode: 0000 [1] SMP CPU 0 Modules linked in: nfs lockd sunrpc af_packet ipv6 cpufreq_ondemand acpi_cpufreq battery floppy nvram sg snd_hda_intel ata_generic snd_pcm_oss snd_mixer_oss snd_pcm i2c_i801 snd_page_alloc e1000 firewire_ohci ata_piix i2c_core sr_mod cdrom sata_sil ahci libata sd_mod scsi_mod ext3 jbd ehci_hcd uhci_hcd Pid: 16353, comm: 10.10.10.1-recl Not tainted 2.6.23-rc3 #1 RIP: 0010:[<ffffffff88240980>] [<ffffffff88240980>] :nfs:encode_open+0x1c0/0x330 RSP: 0018:ffff8100467c5c60 EFLAGS: 00010202 RAX: ffff81000f89b8b8 RBX: 00000000697a6f6d RCX: ffff81000f89b8b8 RDX: 0000000000000004 RSI: 0000000000000004 RDI: ffff8100467c5c80 RBP: ffff8100467c5c80 R08: ffff81000f89bc30 R09: ffff81000f89b83f R10: 0000000000000001 R11: ffffffff881e79e0 R12: ffff81003cbd1808 R13: ffff81000f89b860 R14: ffff81005fc984e0 R15: ffffffff88240af0 FS: 0000000000000000(0000) GS:ffffffff8052a000(0000) knlGS:0000000000000000 CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b CR2: 00002adb9e51a030 CR3: 000000007ea7e000 CR4: 00000000000006e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Process 10.10.10.1-recl (pid: 16353, threadinfo ffff8100467c4000, task ffff8100038ce780) Stack: ffff81004aeb6a40 ffff81003cbd1808 ffff81003cbd1808 ffffffff88240b5d ffff81000f89b8bc ffff81005fc984e8 ffff81000f89bc30 ffff81005fc984e8 0000000300000000 0000000000000000 0000000000000000 ffff81003cbd1800 Call Trace: [<ffffffff88240b5d>] :nfs:nfs4_xdr_enc_open_noattr+0x6d/0x90 [<ffffffff881e74b7>] :sunrpc:rpcauth_wrap_req+0x97/0xf0 [<ffffffff88240af0>] :nfs:nfs4_xdr_enc_open_noattr+0x0/0x90 [<ffffffff881df57a>] :sunrpc:call_transmit+0x18a/0x290 [<ffffffff881e5e7b>] :sunrpc:__rpc_execute+0x6b/0x290 [<ffffffff881dff76>] :sunrpc:rpc_do_run_task+0x76/0xd0 [<ffffffff882373f6>] :nfs:_nfs4_proc_open+0x76/0x230 [<ffffffff88237a2e>] :nfs:nfs4_open_recover_helper+0x5e/0xc0 [<ffffffff88237b74>] :nfs:nfs4_open_recover+0xe4/0x120 [<ffffffff88238e14>] :nfs:nfs4_open_reclaim+0xa4/0xf0 [<ffffffff882413c5>] :nfs:nfs4_reclaim_open_state+0x55/0x1b0 [<ffffffff882417ea>] :nfs:reclaimer+0x2ca/0x390 [<ffffffff88241520>] :nfs:reclaimer+0x0/0x390 [<ffffffff8024e59b>] kthread+0x4b/0x80 [<ffffffff8020cad8>] child_rip+0xa/0x12 [<ffffffff8024e550>] kthread+0x0/0x80 [<ffffffff8020cace>] child_rip+0x0/0x12 Code: 0f 0b eb fe 48 89 ef c7 00 00 00 00 02 be 08 00 00 00 e8 79 RIP [<ffffffff88240980>] :nfs:encode_open+0x1c0/0x330 RSP <ffff8100467c5c60> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2007-08-27 13:57:46 +00:00
delegation_type = delegation->type;
rcu_read_unlock();
opendata->o_arg.u.delegation_type = delegation_type;
status = nfs4_open_recover(opendata, state);
nfs4_opendata_put(opendata);
return status;
}
static int nfs4_do_open_reclaim(struct nfs_open_context *ctx, struct nfs4_state *state)
{
struct nfs_server *server = NFS_SERVER(state->inode);
struct nfs4_exception exception = { };
int err;
do {
err = _nfs4_do_open_reclaim(ctx, state);
if (err != -NFS4ERR_DELAY)
break;
nfs4_handle_exception(server, err, &exception);
} while (exception.retry);
return err;
}
static int nfs4_open_reclaim(struct nfs4_state_owner *sp, struct nfs4_state *state)
{
struct nfs_open_context *ctx;
int ret;
ctx = nfs4_state_find_open_context(state);
if (IS_ERR(ctx))
return PTR_ERR(ctx);
ret = nfs4_do_open_reclaim(ctx, state);
put_nfs_open_context(ctx);
return ret;
}
static int _nfs4_open_delegation_recall(struct nfs_open_context *ctx, struct nfs4_state *state, const nfs4_stateid *stateid)
{
struct nfs4_opendata *opendata;
int ret;
opendata = nfs4_open_recoverdata_alloc(ctx, state);
if (IS_ERR(opendata))
return PTR_ERR(opendata);
opendata->o_arg.claim = NFS4_OPEN_CLAIM_DELEGATE_CUR;
memcpy(opendata->o_arg.u.delegation.data, stateid->data,
sizeof(opendata->o_arg.u.delegation.data));
ret = nfs4_open_recover(opendata, state);
nfs4_opendata_put(opendata);
return ret;
}
int nfs4_open_delegation_recall(struct nfs_open_context *ctx, struct nfs4_state *state, const nfs4_stateid *stateid)
{
struct nfs4_exception exception = { };
struct nfs_server *server = NFS_SERVER(state->inode);
int err;
do {
err = _nfs4_open_delegation_recall(ctx, state, stateid);
switch (err) {
case 0:
return err;
case -NFS4ERR_STALE_CLIENTID:
case -NFS4ERR_STALE_STATEID:
case -NFS4ERR_EXPIRED:
/* Don't recall a delegation if it was lost */
nfs4_schedule_state_recovery(server->nfs_client);
return err;
}
err = nfs4_handle_exception(server, err, &exception);
} while (exception.retry);
return err;
}
static void nfs4_open_confirm_done(struct rpc_task *task, void *calldata)
{
struct nfs4_opendata *data = calldata;
data->rpc_status = task->tk_status;
if (RPC_ASSASSINATED(task))
return;
if (data->rpc_status == 0) {
memcpy(data->o_res.stateid.data, data->c_res.stateid.data,
sizeof(data->o_res.stateid.data));
nfs_confirm_seqid(&data->owner->so_seqid, 0);
renew_lease(data->o_res.server, data->timestamp);
data->rpc_done = 1;
}
}
static void nfs4_open_confirm_release(void *calldata)
{
struct nfs4_opendata *data = calldata;
struct nfs4_state *state = NULL;
/* If this request hasn't been cancelled, do nothing */
if (data->cancelled == 0)
goto out_free;
/* In case of error, no cleanup! */
if (!data->rpc_done)
goto out_free;
state = nfs4_opendata_to_nfs4_state(data);
if (!IS_ERR(state))
nfs4_close_state(&data->path, state, data->o_arg.open_flags);
out_free:
nfs4_opendata_put(data);
}
static const struct rpc_call_ops nfs4_open_confirm_ops = {
.rpc_call_done = nfs4_open_confirm_done,
.rpc_release = nfs4_open_confirm_release,
};
/*
* Note: On error, nfs4_proc_open_confirm will free the struct nfs4_opendata
*/
static int _nfs4_proc_open_confirm(struct nfs4_opendata *data)
{
struct nfs_server *server = NFS_SERVER(data->dir->d_inode);
struct rpc_task *task;
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_CONFIRM],
.rpc_argp = &data->c_arg,
.rpc_resp = &data->c_res,
.rpc_cred = data->owner->so_cred,
};
struct rpc_task_setup task_setup_data = {
.rpc_client = server->client,
.rpc_message = &msg,
.callback_ops = &nfs4_open_confirm_ops,
.callback_data = data,
.workqueue = nfsiod_workqueue,
.flags = RPC_TASK_ASYNC,
};
int status;
kref_get(&data->kref);
data->rpc_done = 0;
data->rpc_status = 0;
data->timestamp = jiffies;
task = rpc_run_task(&task_setup_data);
if (IS_ERR(task))
return PTR_ERR(task);
status = nfs4_wait_for_completion_rpc_task(task);
if (status != 0) {
data->cancelled = 1;
smp_wmb();
} else
status = data->rpc_status;
rpc_put_task(task);
return status;
}
static void nfs4_open_prepare(struct rpc_task *task, void *calldata)
{
struct nfs4_opendata *data = calldata;
struct nfs4_state_owner *sp = data->owner;
if (nfs_wait_on_sequence(data->o_arg.seqid, task) != 0)
return;
/*
* Check if we still need to send an OPEN call, or if we can use
* a delegation instead.
*/
if (data->state != NULL) {
struct nfs_delegation *delegation;
if (can_open_cached(data->state, data->o_arg.open_flags & (FMODE_READ|FMODE_WRITE|O_EXCL)))
goto out_no_action;
rcu_read_lock();
delegation = rcu_dereference(NFS_I(data->state->inode)->delegation);
if (delegation != NULL &&
(delegation->flags & NFS_DELEGATION_NEED_RECLAIM) == 0) {
rcu_read_unlock();
goto out_no_action;
}
rcu_read_unlock();
}
/* Update sequence id. */
data->o_arg.id = sp->so_owner_id.id;
data->o_arg.clientid = sp->so_client->cl_clientid;
if (data->o_arg.claim == NFS4_OPEN_CLAIM_PREVIOUS) {
task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_NOATTR];
nfs_copy_fh(&data->o_res.fh, data->o_arg.fh);
}
data->timestamp = jiffies;
rpc_call_start(task);
return;
out_no_action:
task->tk_action = NULL;
}
static void nfs4_open_done(struct rpc_task *task, void *calldata)
{
struct nfs4_opendata *data = calldata;
data->rpc_status = task->tk_status;
if (RPC_ASSASSINATED(task))
return;
if (task->tk_status == 0) {
switch (data->o_res.f_attr->mode & S_IFMT) {
case S_IFREG:
break;
case S_IFLNK:
data->rpc_status = -ELOOP;
break;
case S_IFDIR:
data->rpc_status = -EISDIR;
break;
default:
data->rpc_status = -ENOTDIR;
}
renew_lease(data->o_res.server, data->timestamp);
if (!(data->o_res.rflags & NFS4_OPEN_RESULT_CONFIRM))
nfs_confirm_seqid(&data->owner->so_seqid, 0);
}
data->rpc_done = 1;
}
static void nfs4_open_release(void *calldata)
{
struct nfs4_opendata *data = calldata;
struct nfs4_state *state = NULL;
/* If this request hasn't been cancelled, do nothing */
if (data->cancelled == 0)
goto out_free;
/* In case of error, no cleanup! */
if (data->rpc_status != 0 || !data->rpc_done)
goto out_free;
/* In case we need an open_confirm, no cleanup! */
if (data->o_res.rflags & NFS4_OPEN_RESULT_CONFIRM)
goto out_free;
state = nfs4_opendata_to_nfs4_state(data);
if (!IS_ERR(state))
nfs4_close_state(&data->path, state, data->o_arg.open_flags);
out_free:
nfs4_opendata_put(data);
}
static const struct rpc_call_ops nfs4_open_ops = {
.rpc_call_prepare = nfs4_open_prepare,
.rpc_call_done = nfs4_open_done,
.rpc_release = nfs4_open_release,
};
/*
* Note: On error, nfs4_proc_open will free the struct nfs4_opendata
*/
static int _nfs4_proc_open(struct nfs4_opendata *data)
{
struct inode *dir = data->dir->d_inode;
struct nfs_server *server = NFS_SERVER(dir);
struct nfs_openargs *o_arg = &data->o_arg;
struct nfs_openres *o_res = &data->o_res;
struct rpc_task *task;
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN],
.rpc_argp = o_arg,
.rpc_resp = o_res,
.rpc_cred = data->owner->so_cred,
};
struct rpc_task_setup task_setup_data = {
.rpc_client = server->client,
.rpc_message = &msg,
.callback_ops = &nfs4_open_ops,
.callback_data = data,
.workqueue = nfsiod_workqueue,
.flags = RPC_TASK_ASYNC,
};
int status;
kref_get(&data->kref);
data->rpc_done = 0;
data->rpc_status = 0;
data->cancelled = 0;
task = rpc_run_task(&task_setup_data);
if (IS_ERR(task))
return PTR_ERR(task);
status = nfs4_wait_for_completion_rpc_task(task);
if (status != 0) {
data->cancelled = 1;
smp_wmb();
} else
status = data->rpc_status;
rpc_put_task(task);
if (status != 0 || !data->rpc_done)
return status;
if (o_res->fh.size == 0)
_nfs4_proc_lookup(dir, o_arg->name, &o_res->fh, o_res->f_attr);
if (o_arg->open_flags & O_CREAT) {
update_changeattr(dir, &o_res->cinfo);
nfs_post_op_update_inode(dir, o_res->dir_attr);
} else
nfs_refresh_inode(dir, o_res->dir_attr);
if(o_res->rflags & NFS4_OPEN_RESULT_CONFIRM) {
status = _nfs4_proc_open_confirm(data);
if (status != 0)
return status;
}
if (!(o_res->f_attr->valid & NFS_ATTR_FATTR))
_nfs4_proc_getattr(server, &o_res->fh, o_res->f_attr);
return 0;
}
static int nfs4_recover_expired_lease(struct nfs_server *server)
{
struct nfs_client *clp = server->nfs_client;
int ret;
for (;;) {
ret = nfs4_wait_clnt_recover(server->client, clp);
if (ret != 0)
return ret;
if (!test_and_clear_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state))
break;
nfs4_schedule_state_recovery(clp);
}
return 0;
}
/*
* OPEN_EXPIRED:
* reclaim state on the server after a network partition.
* Assumes caller holds the appropriate lock
*/
static int _nfs4_open_expired(struct nfs_open_context *ctx, struct nfs4_state *state)
{
struct nfs4_opendata *opendata;
int ret;
opendata = nfs4_open_recoverdata_alloc(ctx, state);
if (IS_ERR(opendata))
return PTR_ERR(opendata);
ret = nfs4_open_recover(opendata, state);
if (ret == -ESTALE)
d_drop(ctx->path.dentry);
nfs4_opendata_put(opendata);
return ret;
}
static inline int nfs4_do_open_expired(struct nfs_open_context *ctx, struct nfs4_state *state)
{
struct nfs_server *server = NFS_SERVER(state->inode);
struct nfs4_exception exception = { };
int err;
do {
err = _nfs4_open_expired(ctx, state);
if (err == -NFS4ERR_DELAY)
nfs4_handle_exception(server, err, &exception);
} while (exception.retry);
return err;
}
static int nfs4_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state)
{
struct nfs_open_context *ctx;
int ret;
ctx = nfs4_state_find_open_context(state);
if (IS_ERR(ctx))
return PTR_ERR(ctx);
ret = nfs4_do_open_expired(ctx, state);
put_nfs_open_context(ctx);
return ret;
}
/*
* on an EXCLUSIVE create, the server should send back a bitmask with FATTR4-*
* fields corresponding to attributes that were used to store the verifier.
* Make sure we clobber those fields in the later setattr call
*/
static inline void nfs4_exclusive_attrset(struct nfs4_opendata *opendata, struct iattr *sattr)
{
if ((opendata->o_res.attrset[1] & FATTR4_WORD1_TIME_ACCESS) &&
!(sattr->ia_valid & ATTR_ATIME_SET))
sattr->ia_valid |= ATTR_ATIME;
if ((opendata->o_res.attrset[1] & FATTR4_WORD1_TIME_MODIFY) &&
!(sattr->ia_valid & ATTR_MTIME_SET))
sattr->ia_valid |= ATTR_MTIME;
}
/*
* Returns a referenced nfs4_state
*/
static int _nfs4_do_open(struct inode *dir, struct path *path, int flags, struct iattr *sattr, struct rpc_cred *cred, struct nfs4_state **res)
{
struct nfs4_state_owner *sp;
struct nfs4_state *state = NULL;
struct nfs_server *server = NFS_SERVER(dir);
struct nfs_client *clp = server->nfs_client;
struct nfs4_opendata *opendata;
int status;
/* Protect against reboot recovery conflicts */
status = -ENOMEM;
if (!(sp = nfs4_get_state_owner(server, cred))) {
dprintk("nfs4_do_open: nfs4_get_state_owner failed!\n");
goto out_err;
}
status = nfs4_recover_expired_lease(server);
if (status != 0)
goto err_put_state_owner;
if (path->dentry->d_inode != NULL)
nfs4_return_incompatible_delegation(path->dentry->d_inode, flags & (FMODE_READ|FMODE_WRITE));
down_read(&clp->cl_sem);
status = -ENOMEM;
opendata = nfs4_opendata_alloc(path, sp, flags, sattr);
if (opendata == NULL)
goto err_release_rwsem;
if (path->dentry->d_inode != NULL)
opendata->state = nfs4_get_open_state(path->dentry->d_inode, sp);
status = _nfs4_proc_open(opendata);
if (status != 0)
goto err_opendata_put;
if (opendata->o_arg.open_flags & O_EXCL)
nfs4_exclusive_attrset(opendata, sattr);
state = nfs4_opendata_to_nfs4_state(opendata);
status = PTR_ERR(state);
if (IS_ERR(state))
goto err_opendata_put;
nfs4_opendata_put(opendata);
nfs4_put_state_owner(sp);
up_read(&clp->cl_sem);
*res = state;
return 0;
err_opendata_put:
nfs4_opendata_put(opendata);
err_release_rwsem:
up_read(&clp->cl_sem);
err_put_state_owner:
nfs4_put_state_owner(sp);
out_err:
*res = NULL;
return status;
}
static struct nfs4_state *nfs4_do_open(struct inode *dir, struct path *path, int flags, struct iattr *sattr, struct rpc_cred *cred)
{
struct nfs4_exception exception = { };
struct nfs4_state *res;
int status;
do {
status = _nfs4_do_open(dir, path, flags, sattr, cred, &res);
if (status == 0)
break;
/* NOTE: BAD_SEQID means the server and client disagree about the
* book-keeping w.r.t. state-changing operations
* (OPEN/CLOSE/LOCK/LOCKU...)
* It is actually a sign of a bug on the client or on the server.
*
* If we receive a BAD_SEQID error in the particular case of
* doing an OPEN, we assume that nfs_increment_open_seqid() will
* have unhashed the old state_owner for us, and that we can
* therefore safely retry using a new one. We should still warn
* the user though...
*/
if (status == -NFS4ERR_BAD_SEQID) {
printk(KERN_WARNING "NFS: v4 server %s "
" returned a bad sequence-id error!\n",
NFS_SERVER(dir)->nfs_client->cl_hostname);
exception.retry = 1;
continue;
}
/*
* BAD_STATEID on OPEN means that the server cancelled our
* state before it received the OPEN_CONFIRM.
* Recover by retrying the request as per the discussion
* on Page 181 of RFC3530.
*/
if (status == -NFS4ERR_BAD_STATEID) {
exception.retry = 1;
continue;
}
if (status == -EAGAIN) {
/* We must have found a delegation */
exception.retry = 1;
continue;
}
res = ERR_PTR(nfs4_handle_exception(NFS_SERVER(dir),
status, &exception));
} while (exception.retry);
return res;
}
static int _nfs4_do_setattr(struct inode *inode, struct nfs_fattr *fattr,
struct iattr *sattr, struct nfs4_state *state)
{
struct nfs_server *server = NFS_SERVER(inode);
struct nfs_setattrargs arg = {
.fh = NFS_FH(inode),
.iap = sattr,
.server = server,
.bitmask = server->attr_bitmask,
};
struct nfs_setattrres res = {
.fattr = fattr,
.server = server,
};
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETATTR],
.rpc_argp = &arg,
.rpc_resp = &res,
};
unsigned long timestamp = jiffies;
int status;
nfs_fattr_init(fattr);
if (nfs4_copy_delegation_stateid(&arg.stateid, inode)) {
/* Use that stateid */
} else if (state != NULL) {
msg.rpc_cred = state->owner->so_cred;
nfs4_copy_stateid(&arg.stateid, state, current->files);
} else
memcpy(&arg.stateid, &zero_stateid, sizeof(arg.stateid));
status = rpc_call_sync(server->client, &msg, 0);
if (status == 0 && state != NULL)
renew_lease(server, timestamp);
return status;
}
static int nfs4_do_setattr(struct inode *inode, struct nfs_fattr *fattr,
struct iattr *sattr, struct nfs4_state *state)
{
struct nfs_server *server = NFS_SERVER(inode);
struct nfs4_exception exception = { };
int err;
do {
err = nfs4_handle_exception(server,
_nfs4_do_setattr(inode, fattr, sattr, state),
&exception);
} while (exception.retry);
return err;
}
struct nfs4_closedata {
struct path path;
struct inode *inode;
struct nfs4_state *state;
struct nfs_closeargs arg;
struct nfs_closeres res;
struct nfs_fattr fattr;
unsigned long timestamp;
};
static void nfs4_free_closedata(void *data)
{
struct nfs4_closedata *calldata = data;
struct nfs4_state_owner *sp = calldata->state->owner;
nfs4_put_open_state(calldata->state);
nfs_free_seqid(calldata->arg.seqid);
nfs4_put_state_owner(sp);
path_put(&calldata->path);
kfree(calldata);
}
static void nfs4_close_done(struct rpc_task *task, void *data)
{
struct nfs4_closedata *calldata = data;
struct nfs4_state *state = calldata->state;
struct nfs_server *server = NFS_SERVER(calldata->inode);
if (RPC_ASSASSINATED(task))
return;
/* hmm. we are done with the inode, and in the process of freeing
* the state_owner. we keep this around to process errors
*/
switch (task->tk_status) {
case 0:
nfs_set_open_stateid(state, &calldata->res.stateid, 0);
renew_lease(server, calldata->timestamp);
break;
case -NFS4ERR_STALE_STATEID:
case -NFS4ERR_EXPIRED:
break;
default:
if (nfs4_async_handle_error(task, server) == -EAGAIN) {
rpc_restart_call(task);
return;
}
}
nfs_refresh_inode(calldata->inode, calldata->res.fattr);
}
static void nfs4_close_prepare(struct rpc_task *task, void *data)
{
struct nfs4_closedata *calldata = data;
struct nfs4_state *state = calldata->state;
int clear_rd, clear_wr, clear_rdwr;
if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
return;
clear_rd = clear_wr = clear_rdwr = 0;
spin_lock(&state->owner->so_lock);
/* Calculate the change in open mode */
if (state->n_rdwr == 0) {
if (state->n_rdonly == 0) {
clear_rd |= test_and_clear_bit(NFS_O_RDONLY_STATE, &state->flags);
clear_rdwr |= test_and_clear_bit(NFS_O_RDWR_STATE, &state->flags);
}
if (state->n_wronly == 0) {
clear_wr |= test_and_clear_bit(NFS_O_WRONLY_STATE, &state->flags);
clear_rdwr |= test_and_clear_bit(NFS_O_RDWR_STATE, &state->flags);
}
}
spin_unlock(&state->owner->so_lock);
if (!clear_rd && !clear_wr && !clear_rdwr) {
/* Note: exit _without_ calling nfs4_close_done */
task->tk_action = NULL;
return;
}
nfs_fattr_init(calldata->res.fattr);
if (test_bit(NFS_O_RDONLY_STATE, &state->flags) != 0) {
task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_DOWNGRADE];
calldata->arg.open_flags = FMODE_READ;
} else if (test_bit(NFS_O_WRONLY_STATE, &state->flags) != 0) {
task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_DOWNGRADE];
calldata->arg.open_flags = FMODE_WRITE;
}
calldata->timestamp = jiffies;
rpc_call_start(task);
}
static const struct rpc_call_ops nfs4_close_ops = {
.rpc_call_prepare = nfs4_close_prepare,
.rpc_call_done = nfs4_close_done,
.rpc_release = nfs4_free_closedata,
};
/*
* It is possible for data to be read/written from a mem-mapped file
* after the sys_close call (which hits the vfs layer as a flush).
* This means that we can't safely call nfsv4 close on a file until
* the inode is cleared. This in turn means that we are not good
* NFSv4 citizens - we do not indicate to the server to update the file's
* share state even when we are done with one of the three share
* stateid's in the inode.
*
* NOTE: Caller must be holding the sp->so_owner semaphore!
*/
int nfs4_do_close(struct path *path, struct nfs4_state *state, int wait)
{
struct nfs_server *server = NFS_SERVER(state->inode);
struct nfs4_closedata *calldata;
struct nfs4_state_owner *sp = state->owner;
struct rpc_task *task;
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CLOSE],
.rpc_cred = state->owner->so_cred,
};
struct rpc_task_setup task_setup_data = {
.rpc_client = server->client,
.rpc_message = &msg,
.callback_ops = &nfs4_close_ops,
.workqueue = nfsiod_workqueue,
.flags = RPC_TASK_ASYNC,
};
int status = -ENOMEM;
calldata = kmalloc(sizeof(*calldata), GFP_KERNEL);
if (calldata == NULL)
goto out;
calldata->inode = state->inode;
calldata->state = state;
calldata->arg.fh = NFS_FH(state->inode);
calldata->arg.stateid = &state->open_stateid;
/* Serialization for the sequence id */
calldata->arg.seqid = nfs_alloc_seqid(&state->owner->so_seqid);
if (calldata->arg.seqid == NULL)
goto out_free_calldata;
calldata->arg.bitmask = server->attr_bitmask;
calldata->res.fattr = &calldata->fattr;
calldata->res.seqid = calldata->arg.seqid;
calldata->res.server = server;
calldata->path.mnt = mntget(path->mnt);
calldata->path.dentry = dget(path->dentry);
msg.rpc_argp = &calldata->arg,
msg.rpc_resp = &calldata->res,
task_setup_data.callback_data = calldata;
task = rpc_run_task(&task_setup_data);
if (IS_ERR(task))
return PTR_ERR(task);
status = 0;
if (wait)
status = rpc_wait_for_completion_task(task);
rpc_put_task(task);
return status;
out_free_calldata:
kfree(calldata);
out:
nfs4_put_open_state(state);
nfs4_put_state_owner(sp);
return status;
}
static int nfs4_intent_set_file(struct nameidata *nd, struct path *path, struct nfs4_state *state)
{
struct file *filp;
int ret;
/* If the open_intent is for execute, we have an extra check to make */
if (nd->intent.open.flags & FMODE_EXEC) {
ret = nfs_may_open(state->inode,
state->owner->so_cred,
nd->intent.open.flags);
if (ret < 0)
goto out_close;
}
filp = lookup_instantiate_filp(nd, path->dentry, NULL);
if (!IS_ERR(filp)) {
struct nfs_open_context *ctx;
ctx = nfs_file_open_context(filp);
ctx->state = state;
return 0;
}
ret = PTR_ERR(filp);
out_close:
nfs4_close_sync(path, state, nd->intent.open.flags);
return ret;
}
struct dentry *
nfs4_atomic_open(struct inode *dir, struct dentry *dentry, struct nameidata *nd)
{
struct path path = {
.mnt = nd->path.mnt,
.dentry = dentry,
};
struct dentry *parent;
struct iattr attr;
struct rpc_cred *cred;
struct nfs4_state *state;
struct dentry *res;
if (nd->flags & LOOKUP_CREATE) {
attr.ia_mode = nd->intent.open.create_mode;
attr.ia_valid = ATTR_MODE;
if (!IS_POSIXACL(dir))
attr.ia_mode &= ~current->fs->umask;
} else {
attr.ia_valid = 0;
BUG_ON(nd->intent.open.flags & O_CREAT);
}
cred = rpc_lookup_cred();
if (IS_ERR(cred))
return (struct dentry *)cred;
parent = dentry->d_parent;
/* Protect against concurrent sillydeletes */
nfs_block_sillyrename(parent);
state = nfs4_do_open(dir, &path, nd->intent.open.flags, &attr, cred);
put_rpccred(cred);
if (IS_ERR(state)) {
if (PTR_ERR(state) == -ENOENT) {
d_add(dentry, NULL);
nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
}
nfs_unblock_sillyrename(parent);
return (struct dentry *)state;
}
res = d_add_unique(dentry, igrab(state->inode));
if (res != NULL)
path.dentry = res;
nfs_set_verifier(path.dentry, nfs_save_change_attribute(dir));
nfs_unblock_sillyrename(parent);
nfs4_intent_set_file(nd, &path, state);
return res;
}
int
nfs4_open_revalidate(struct inode *dir, struct dentry *dentry, int openflags, struct nameidata *nd)
{
struct path path = {
.mnt = nd->path.mnt,
.dentry = dentry,
};
struct rpc_cred *cred;
struct nfs4_state *state;
cred = rpc_lookup_cred();
if (IS_ERR(cred))
return PTR_ERR(cred);
state = nfs4_do_open(dir, &path, openflags, NULL, cred);
put_rpccred(cred);
if (IS_ERR(state)) {
switch (PTR_ERR(state)) {
case -EPERM:
case -EACCES:
case -EDQUOT:
case -ENOSPC:
case -EROFS:
lookup_instantiate_filp(nd, (struct dentry *)state, NULL);
return 1;
default:
goto out_drop;
}
}
if (state->inode == dentry->d_inode) {
nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
nfs4_intent_set_file(nd, &path, state);
return 1;
}
nfs4_close_sync(&path, state, openflags);
out_drop:
d_drop(dentry);
return 0;
}
static int _nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *fhandle)
{
struct nfs4_server_caps_res res = {};
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SERVER_CAPS],
.rpc_argp = fhandle,
.rpc_resp = &res,
};
int status;
status = rpc_call_sync(server->client, &msg, 0);
if (status == 0) {
memcpy(server->attr_bitmask, res.attr_bitmask, sizeof(server->attr_bitmask));
if (res.attr_bitmask[0] & FATTR4_WORD0_ACL)
server->caps |= NFS_CAP_ACLS;
if (res.has_links != 0)
server->caps |= NFS_CAP_HARDLINKS;
if (res.has_symlinks != 0)
server->caps |= NFS_CAP_SYMLINKS;
server->acl_bitmask = res.acl_bitmask;
}
return status;
}
int nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *fhandle)
{
struct nfs4_exception exception = { };
int err;
do {
err = nfs4_handle_exception(server,
_nfs4_server_capabilities(server, fhandle),
&exception);
} while (exception.retry);
return err;
}
static int _nfs4_lookup_root(struct nfs_server *server, struct nfs_fh *fhandle,
struct nfs_fsinfo *info)
{
struct nfs4_lookup_root_arg args = {
.bitmask = nfs4_fattr_bitmap,
};
struct nfs4_lookup_res res = {
.server = server,
.fattr = info->fattr,
.fh = fhandle,
};
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOOKUP_ROOT],
.rpc_argp = &args,
.rpc_resp = &res,
};
nfs_fattr_init(info->fattr);
return rpc_call_sync(server->client, &msg, 0);
}
static int nfs4_lookup_root(struct nfs_server *server, struct nfs_fh *fhandle,
struct nfs_fsinfo *info)
{
struct nfs4_exception exception = { };
int err;
do {
err = nfs4_handle_exception(server,
_nfs4_lookup_root(server, fhandle, info),
&exception);
} while (exception.retry);
return err;
}
NFS: Share NFS superblocks per-protocol per-server per-FSID The attached patch makes NFS share superblocks between mounts from the same server and FSID over the same protocol. It does this by creating each superblock with a false root and returning the real root dentry in the vfsmount presented by get_sb(). The root dentry set starts off as an anonymous dentry if we don't already have the dentry for its inode, otherwise it simply returns the dentry we already have. We may thus end up with several trees of dentries in the superblock, and if at some later point one of anonymous tree roots is discovered by normal filesystem activity to be located in another tree within the superblock, the anonymous root is named and materialises attached to the second tree at the appropriate point. Why do it this way? Why not pass an extra argument to the mount() syscall to indicate the subpath and then pathwalk from the server root to the desired directory? You can't guarantee this will work for two reasons: (1) The root and intervening nodes may not be accessible to the client. With NFS2 and NFS3, for instance, mountd is called on the server to get the filehandle for the tip of a path. mountd won't give us handles for anything we don't have permission to access, and so we can't set up NFS inodes for such nodes, and so can't easily set up dentries (we'd have to have ghost inodes or something). With this patch we don't actually create dentries until we get handles from the server that we can use to set up their inodes, and we don't actually bind them into the tree until we know for sure where they go. (2) Inaccessible symbolic links. If we're asked to mount two exports from the server, eg: mount warthog:/warthog/aaa/xxx /mmm mount warthog:/warthog/bbb/yyy /nnn We may not be able to access anything nearer the root than xxx and yyy, but we may find out later that /mmm/www/yyy, say, is actually the same directory as the one mounted on /nnn. What we might then find out, for example, is that /warthog/bbb was actually a symbolic link to /warthog/aaa/xxx/www, but we can't actually determine that by talking to the server until /warthog is made available by NFS. This would lead to having constructed an errneous dentry tree which we can't easily fix. We can end up with a dentry marked as a directory when it should actually be a symlink, or we could end up with an apparently hardlinked directory. With this patch we need not make assumptions about the type of a dentry for which we can't retrieve information, nor need we assume we know its place in the grand scheme of things until we actually see that place. This patch reduces the possibility of aliasing in the inode and page caches for inodes that may be accessed by more than one NFS export. It also reduces the number of superblocks required for NFS where there are many NFS exports being used from a server (home directory server + autofs for example). This in turn makes it simpler to do local caching of network filesystems, as it can then be guaranteed that there won't be links from multiple inodes in separate superblocks to the same cache file. Obviously, cache aliasing between different levels of NFS protocol could still be a problem, but at least that gives us another key to use when indexing the cache. This patch makes the following changes: (1) The server record construction/destruction has been abstracted out into its own set of functions to make things easier to get right. These have been moved into fs/nfs/client.c. All the code in fs/nfs/client.c has to do with the management of connections to servers, and doesn't touch superblocks in any way; the remaining code in fs/nfs/super.c has to do with VFS superblock management. (2) The sequence of events undertaken by NFS mount is now reordered: (a) A volume representation (struct nfs_server) is allocated. (b) A server representation (struct nfs_client) is acquired. This may be allocated or shared, and is keyed on server address, port and NFS version. (c) If allocated, the client representation is initialised. The state member variable of nfs_client is used to prevent a race during initialisation from two mounts. (d) For NFS4 a simple pathwalk is performed, walking from FH to FH to find the root filehandle for the mount (fs/nfs/getroot.c). For NFS2/3 we are given the root FH in advance. (e) The volume FSID is probed for on the root FH. (f) The volume representation is initialised from the FSINFO record retrieved on the root FH. (g) sget() is called to acquire a superblock. This may be allocated or shared, keyed on client pointer and FSID. (h) If allocated, the superblock is initialised. (i) If the superblock is shared, then the new nfs_server record is discarded. (j) The root dentry for this mount is looked up from the root FH. (k) The root dentry for this mount is assigned to the vfsmount. (3) nfs_readdir_lookup() creates dentries for each of the entries readdir() returns; this function now attaches disconnected trees from alternate roots that happen to be discovered attached to a directory being read (in the same way nfs_lookup() is made to do for lookup ops). The new d_materialise_unique() function is now used to do this, thus permitting the whole thing to be done under one set of locks, and thus avoiding any race between mount and lookup operations on the same directory. (4) The client management code uses a new debug facility: NFSDBG_CLIENT which is set by echoing 1024 to /proc/net/sunrpc/nfs_debug. (5) Clone mounts are now called xdev mounts. (6) Use the dentry passed to the statfs() op as the handle for retrieving fs statistics rather than the root dentry of the superblock (which is now a dummy). Signed-Off-By: David Howells <dhowells@redhat.com> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2006-08-23 00:06:13 +00:00
/*
* get the file handle for the "/" directory on the server
*/
static int nfs4_proc_get_root(struct nfs_server *server, struct nfs_fh *fhandle,
NFS: Share NFS superblocks per-protocol per-server per-FSID The attached patch makes NFS share superblocks between mounts from the same server and FSID over the same protocol. It does this by creating each superblock with a false root and returning the real root dentry in the vfsmount presented by get_sb(). The root dentry set starts off as an anonymous dentry if we don't already have the dentry for its inode, otherwise it simply returns the dentry we already have. We may thus end up with several trees of dentries in the superblock, and if at some later point one of anonymous tree roots is discovered by normal filesystem activity to be located in another tree within the superblock, the anonymous root is named and materialises attached to the second tree at the appropriate point. Why do it this way? Why not pass an extra argument to the mount() syscall to indicate the subpath and then pathwalk from the server root to the desired directory? You can't guarantee this will work for two reasons: (1) The root and intervening nodes may not be accessible to the client. With NFS2 and NFS3, for instance, mountd is called on the server to get the filehandle for the tip of a path. mountd won't give us handles for anything we don't have permission to access, and so we can't set up NFS inodes for such nodes, and so can't easily set up dentries (we'd have to have ghost inodes or something). With this patch we don't actually create dentries until we get handles from the server that we can use to set up their inodes, and we don't actually bind them into the tree until we know for sure where they go. (2) Inaccessible symbolic links. If we're asked to mount two exports from the server, eg: mount warthog:/warthog/aaa/xxx /mmm mount warthog:/warthog/bbb/yyy /nnn We may not be able to access anything nearer the root than xxx and yyy, but we may find out later that /mmm/www/yyy, say, is actually the same directory as the one mounted on /nnn. What we might then find out, for example, is that /warthog/bbb was actually a symbolic link to /warthog/aaa/xxx/www, but we can't actually determine that by talking to the server until /warthog is made available by NFS. This would lead to having constructed an errneous dentry tree which we can't easily fix. We can end up with a dentry marked as a directory when it should actually be a symlink, or we could end up with an apparently hardlinked directory. With this patch we need not make assumptions about the type of a dentry for which we can't retrieve information, nor need we assume we know its place in the grand scheme of things until we actually see that place. This patch reduces the possibility of aliasing in the inode and page caches for inodes that may be accessed by more than one NFS export. It also reduces the number of superblocks required for NFS where there are many NFS exports being used from a server (home directory server + autofs for example). This in turn makes it simpler to do local caching of network filesystems, as it can then be guaranteed that there won't be links from multiple inodes in separate superblocks to the same cache file. Obviously, cache aliasing between different levels of NFS protocol could still be a problem, but at least that gives us another key to use when indexing the cache. This patch makes the following changes: (1) The server record construction/destruction has been abstracted out into its own set of functions to make things easier to get right. These have been moved into fs/nfs/client.c. All the code in fs/nfs/client.c has to do with the management of connections to servers, and doesn't touch superblocks in any way; the remaining code in fs/nfs/super.c has to do with VFS superblock management. (2) The sequence of events undertaken by NFS mount is now reordered: (a) A volume representation (struct nfs_server) is allocated. (b) A server representation (struct nfs_client) is acquired. This may be allocated or shared, and is keyed on server address, port and NFS version. (c) If allocated, the client representation is initialised. The state member variable of nfs_client is used to prevent a race during initialisation from two mounts. (d) For NFS4 a simple pathwalk is performed, walking from FH to FH to find the root filehandle for the mount (fs/nfs/getroot.c). For NFS2/3 we are given the root FH in advance. (e) The volume FSID is probed for on the root FH. (f) The volume representation is initialised from the FSINFO record retrieved on the root FH. (g) sget() is called to acquire a superblock. This may be allocated or shared, keyed on client pointer and FSID. (h) If allocated, the superblock is initialised. (i) If the superblock is shared, then the new nfs_server record is discarded. (j) The root dentry for this mount is looked up from the root FH. (k) The root dentry for this mount is assigned to the vfsmount. (3) nfs_readdir_lookup() creates dentries for each of the entries readdir() returns; this function now attaches disconnected trees from alternate roots that happen to be discovered attached to a directory being read (in the same way nfs_lookup() is made to do for lookup ops). The new d_materialise_unique() function is now used to do this, thus permitting the whole thing to be done under one set of locks, and thus avoiding any race between mount and lookup operations on the same directory. (4) The client management code uses a new debug facility: NFSDBG_CLIENT which is set by echoing 1024 to /proc/net/sunrpc/nfs_debug. (5) Clone mounts are now called xdev mounts. (6) Use the dentry passed to the statfs() op as the handle for retrieving fs statistics rather than the root dentry of the superblock (which is now a dummy). Signed-Off-By: David Howells <dhowells@redhat.com> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2006-08-23 00:06:13 +00:00
struct nfs_fsinfo *info)
{
int status;
status = nfs4_lookup_root(server, fhandle, info);
if (status == 0)
status = nfs4_server_capabilities(server, fhandle);
if (status == 0)
status = nfs4_do_fsinfo(server, fhandle, info);
return nfs4_map_errors(status);
}
/*
* Get locations and (maybe) other attributes of a referral.
* Note that we'll actually follow the referral later when
* we detect fsid mismatch in inode revalidation
*/
static int nfs4_get_referral(struct inode *dir, const struct qstr *name, struct nfs_fattr *fattr, struct nfs_fh *fhandle)
{
int status = -ENOMEM;
struct page *page = NULL;
struct nfs4_fs_locations *locations = NULL;
page = alloc_page(GFP_KERNEL);
if (page == NULL)
goto out;
locations = kmalloc(sizeof(struct nfs4_fs_locations), GFP_KERNEL);
if (locations == NULL)
goto out;
status = nfs4_proc_fs_locations(dir, name, locations, page);
if (status != 0)
goto out;
/* Make sure server returned a different fsid for the referral */
if (nfs_fsid_equal(&NFS_SERVER(dir)->fsid, &locations->fattr.fsid)) {
dprintk("%s: server did not return a different fsid for a referral at %s\n", __func__, name->name);
status = -EIO;
goto out;
}
memcpy(fattr, &locations->fattr, sizeof(struct nfs_fattr));
fattr->valid |= NFS_ATTR_FATTR_V4_REFERRAL;
if (!fattr->mode)
fattr->mode = S_IFDIR;
memset(fhandle, 0, sizeof(struct nfs_fh));
out:
if (page)
__free_page(page);
if (locations)
kfree(locations);
return status;
}
static int _nfs4_proc_getattr(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fattr *fattr)
{
struct nfs4_getattr_arg args = {
.fh = fhandle,
.bitmask = server->attr_bitmask,
};
struct nfs4_getattr_res res = {
.fattr = fattr,
.server = server,
};
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETATTR],
.rpc_argp = &args,
.rpc_resp = &res,
};
nfs_fattr_init(fattr);
return rpc_call_sync(server->client, &msg, 0);
}
static int nfs4_proc_getattr(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fattr *fattr)
{
struct nfs4_exception exception = { };
int err;
do {
err = nfs4_handle_exception(server,
_nfs4_proc_getattr(server, fhandle, fattr),
&exception);
} while (exception.retry);
return err;
}
/*
* The file is not closed if it is opened due to the a request to change
* the size of the file. The open call will not be needed once the
* VFS layer lookup-intents are implemented.
*
* Close is called when the inode is destroyed.
* If we haven't opened the file for O_WRONLY, we
* need to in the size_change case to obtain a stateid.
*
* Got race?
* Because OPEN is always done by name in nfsv4, it is
* possible that we opened a different file by the same
* name. We can recognize this race condition, but we
* can't do anything about it besides returning an error.
*
* This will be fixed with VFS changes (lookup-intent).
*/
static int
nfs4_proc_setattr(struct dentry *dentry, struct nfs_fattr *fattr,
struct iattr *sattr)
{
struct rpc_cred *cred;
struct inode *inode = dentry->d_inode;
struct nfs_open_context *ctx;
struct nfs4_state *state = NULL;
int status;
nfs_fattr_init(fattr);
cred = rpc_lookup_cred();
if (IS_ERR(cred))
return PTR_ERR(cred);
/* Search for an existing open(O_WRITE) file */
ctx = nfs_find_open_context(inode, cred, FMODE_WRITE);
if (ctx != NULL)
state = ctx->state;
status = nfs4_do_setattr(inode, fattr, sattr, state);
if (status == 0)
nfs_setattr_update_inode(inode, sattr);
if (ctx != NULL)
put_nfs_open_context(ctx);
put_rpccred(cred);
return status;
}
static int _nfs4_proc_lookupfh(struct nfs_server *server, const struct nfs_fh *dirfh,
const struct qstr *name, struct nfs_fh *fhandle,
struct nfs_fattr *fattr)
{
int status;
struct nfs4_lookup_arg args = {
.bitmask = server->attr_bitmask,
.dir_fh = dirfh,
.name = name,
};
struct nfs4_lookup_res res = {
.server = server,
.fattr = fattr,
.fh = fhandle,
};
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOOKUP],
.rpc_argp = &args,
.rpc_resp = &res,
};
nfs_fattr_init(fattr);
dprintk("NFS call lookupfh %s\n", name->name);
status = rpc_call_sync(server->client, &msg, 0);
dprintk("NFS reply lookupfh: %d\n", status);
return status;
}
static int nfs4_proc_lookupfh(struct nfs_server *server, struct nfs_fh *dirfh,
struct qstr *name, struct nfs_fh *fhandle,
struct nfs_fattr *fattr)
{
struct nfs4_exception exception = { };
int err;
do {
err = _nfs4_proc_lookupfh(server, dirfh, name, fhandle, fattr);
/* FIXME: !!!! */
if (err == -NFS4ERR_MOVED) {
err = -EREMOTE;
break;
}
err = nfs4_handle_exception(server, err, &exception);
} while (exception.retry);
return err;
}
static int _nfs4_proc_lookup(struct inode *dir, const struct qstr *name,
struct nfs_fh *fhandle, struct nfs_fattr *fattr)
{
int status;
dprintk("NFS call lookup %s\n", name->name);
status = _nfs4_proc_lookupfh(NFS_SERVER(dir), NFS_FH(dir), name, fhandle, fattr);
if (status == -NFS4ERR_MOVED)
status = nfs4_get_referral(dir, name, fattr, fhandle);
dprintk("NFS reply lookup: %d\n", status);
return status;
}
static int nfs4_proc_lookup(struct inode *dir, struct qstr *name, struct nfs_fh *fhandle, struct nfs_fattr *fattr)
{
struct nfs4_exception exception = { };
int err;
do {
err = nfs4_handle_exception(NFS_SERVER(dir),
_nfs4_proc_lookup(dir, name, fhandle, fattr),
&exception);
} while (exception.retry);
return err;
}
static int _nfs4_proc_access(struct inode *inode, struct nfs_access_entry *entry)
{
struct nfs_server *server = NFS_SERVER(inode);
struct nfs_fattr fattr;
struct nfs4_accessargs args = {
.fh = NFS_FH(inode),
.bitmask = server->attr_bitmask,
};
struct nfs4_accessres res = {
.server = server,
.fattr = &fattr,
};
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_ACCESS],
.rpc_argp = &args,
.rpc_resp = &res,
.rpc_cred = entry->cred,
};
int mode = entry->mask;
int status;
/*
* Determine which access bits we want to ask for...
*/
if (mode & MAY_READ)
args.access |= NFS4_ACCESS_READ;
if (S_ISDIR(inode->i_mode)) {
if (mode & MAY_WRITE)
args.access |= NFS4_ACCESS_MODIFY | NFS4_ACCESS_EXTEND | NFS4_ACCESS_DELETE;
if (mode & MAY_EXEC)
args.access |= NFS4_ACCESS_LOOKUP;
} else {
if (mode & MAY_WRITE)
args.access |= NFS4_ACCESS_MODIFY | NFS4_ACCESS_EXTEND;
if (mode & MAY_EXEC)
args.access |= NFS4_ACCESS_EXECUTE;
}
nfs_fattr_init(&fattr);
status = rpc_call_sync(NFS_CLIENT(inode), &msg, 0);
if (!status) {
entry->mask = 0;
if (res.access & NFS4_ACCESS_READ)
entry->mask |= MAY_READ;
if (res.access & (NFS4_ACCESS_MODIFY | NFS4_ACCESS_EXTEND | NFS4_ACCESS_DELETE))
entry->mask |= MAY_WRITE;
if (res.access & (NFS4_ACCESS_LOOKUP|NFS4_ACCESS_EXECUTE))
entry->mask |= MAY_EXEC;
nfs_refresh_inode(inode, &fattr);
}
return status;
}
static int nfs4_proc_access(struct inode *inode, struct nfs_access_entry *entry)
{
struct nfs4_exception exception = { };
int err;
do {
err = nfs4_handle_exception(NFS_SERVER(inode),
_nfs4_proc_access(inode, entry),
&exception);
} while (exception.retry);
return err;
}
/*
* TODO: For the time being, we don't try to get any attributes
* along with any of the zero-copy operations READ, READDIR,
* READLINK, WRITE.
*
* In the case of the first three, we want to put the GETATTR
* after the read-type operation -- this is because it is hard
* to predict the length of a GETATTR response in v4, and thus
* align the READ data correctly. This means that the GETATTR
* may end up partially falling into the page cache, and we should
* shift it into the 'tail' of the xdr_buf before processing.
* To do this efficiently, we need to know the total length
* of data received, which doesn't seem to be available outside
* of the RPC layer.
*
* In the case of WRITE, we also want to put the GETATTR after
* the operation -- in this case because we want to make sure
* we get the post-operation mtime and size. This means that
* we can't use xdr_encode_pages() as written: we need a variant
* of it which would leave room in the 'tail' iovec.
*
* Both of these changes to the XDR layer would in fact be quite
* minor, but I decided to leave them for a subsequent patch.
*/
static int _nfs4_proc_readlink(struct inode *inode, struct page *page,
unsigned int pgbase, unsigned int pglen)
{
struct nfs4_readlink args = {
.fh = NFS_FH(inode),
.pgbase = pgbase,
.pglen = pglen,
.pages = &page,
};
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READLINK],
.rpc_argp = &args,
.rpc_resp = NULL,
};
return rpc_call_sync(NFS_CLIENT(inode), &msg, 0);
}
static int nfs4_proc_readlink(struct inode *inode, struct page *page,
unsigned int pgbase, unsigned int pglen)
{
struct nfs4_exception exception = { };
int err;
do {
err = nfs4_handle_exception(NFS_SERVER(inode),
_nfs4_proc_readlink(inode, page, pgbase, pglen),
&exception);
} while (exception.retry);
return err;
}
/*
* Got race?
* We will need to arrange for the VFS layer to provide an atomic open.
* Until then, this create/open method is prone to inefficiency and race
* conditions due to the lookup, create, and open VFS calls from sys_open()
* placed on the wire.
*
* Given the above sorry state of affairs, I'm simply sending an OPEN.
* The file will be opened again in the subsequent VFS open call
* (nfs4_proc_file_open).
*
* The open for read will just hang around to be used by any process that
* opens the file O_RDONLY. This will all be resolved with the VFS changes.
*/
static int
nfs4_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
int flags, struct nameidata *nd)
{
struct path path = {
.mnt = nd->path.mnt,
.dentry = dentry,
};
struct nfs4_state *state;
struct rpc_cred *cred;
int status = 0;
cred = rpc_lookup_cred();
if (IS_ERR(cred)) {
status = PTR_ERR(cred);
goto out;
}
state = nfs4_do_open(dir, &path, flags, sattr, cred);
put_rpccred(cred);
d_drop(dentry);
if (IS_ERR(state)) {
status = PTR_ERR(state);
goto out;
}
d_add(dentry, igrab(state->inode));
nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
if (flags & O_EXCL) {
struct nfs_fattr fattr;
status = nfs4_do_setattr(state->inode, &fattr, sattr, state);
if (status == 0)
nfs_setattr_update_inode(state->inode, sattr);
nfs_post_op_update_inode(state->inode, &fattr);
}
if (status == 0 && (nd->flags & LOOKUP_OPEN) != 0)
status = nfs4_intent_set_file(nd, &path, state);
else
nfs4_close_sync(&path, state, flags);
out:
return status;
}
static int _nfs4_proc_remove(struct inode *dir, struct qstr *name)
{
struct nfs_server *server = NFS_SERVER(dir);
struct nfs_removeargs args = {
.fh = NFS_FH(dir),
.name.len = name->len,
.name.name = name->name,
.bitmask = server->attr_bitmask,
};
struct nfs_removeres res = {
.server = server,
};
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_REMOVE],
.rpc_argp = &args,
.rpc_resp = &res,
};
int status;
nfs_fattr_init(&res.dir_attr);
status = rpc_call_sync(server->client, &msg, 0);
if (status == 0) {
update_changeattr(dir, &res.cinfo);
nfs_post_op_update_inode(dir, &res.dir_attr);
}
return status;
}
static int nfs4_proc_remove(struct inode *dir, struct qstr *name)
{
struct nfs4_exception exception = { };
int err;
do {
err = nfs4_handle_exception(NFS_SERVER(dir),
_nfs4_proc_remove(dir, name),
&exception);
} while (exception.retry);
return err;
}
static void nfs4_proc_unlink_setup(struct rpc_message *msg, struct inode *dir)
{
struct nfs_server *server = NFS_SERVER(dir);
struct nfs_removeargs *args = msg->rpc_argp;
struct nfs_removeres *res = msg->rpc_resp;
args->bitmask = server->attr_bitmask;
res->server = server;
msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_REMOVE];
}
static int nfs4_proc_unlink_done(struct rpc_task *task, struct inode *dir)
{
struct nfs_removeres *res = task->tk_msg.rpc_resp;
if (nfs4_async_handle_error(task, res->server) == -EAGAIN)
return 0;
update_changeattr(dir, &res->cinfo);
nfs_post_op_update_inode(dir, &res->dir_attr);
return 1;
}
static int _nfs4_proc_rename(struct inode *old_dir, struct qstr *old_name,
struct inode *new_dir, struct qstr *new_name)
{
struct nfs_server *server = NFS_SERVER(old_dir);
struct nfs4_rename_arg arg = {
.old_dir = NFS_FH(old_dir),
.new_dir = NFS_FH(new_dir),
.old_name = old_name,
.new_name = new_name,
.bitmask = server->attr_bitmask,
};
struct nfs_fattr old_fattr, new_fattr;
struct nfs4_rename_res res = {
.server = server,
.old_fattr = &old_fattr,
.new_fattr = &new_fattr,
};
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RENAME],
.rpc_argp = &arg,
.rpc_resp = &res,
};
int status;
nfs_fattr_init(res.old_fattr);
nfs_fattr_init(res.new_fattr);
status = rpc_call_sync(server->client, &msg, 0);
if (!status) {
update_changeattr(old_dir, &res.old_cinfo);
nfs_post_op_update_inode(old_dir, res.old_fattr);
update_changeattr(new_dir, &res.new_cinfo);
nfs_post_op_update_inode(new_dir, res.new_fattr);
}
return status;
}
static int nfs4_proc_rename(struct inode *old_dir, struct qstr *old_name,
struct inode *new_dir, struct qstr *new_name)
{
struct nfs4_exception exception = { };
int err;
do {
err = nfs4_handle_exception(NFS_SERVER(old_dir),
_nfs4_proc_rename(old_dir, old_name,
new_dir, new_name),
&exception);
} while (exception.retry);
return err;
}
static int _nfs4_proc_link(struct inode *inode, struct inode *dir, struct qstr *name)
{
struct nfs_server *server = NFS_SERVER(inode);
struct nfs4_link_arg arg = {
.fh = NFS_FH(inode),
.dir_fh = NFS_FH(dir),
.name = name,
.bitmask = server->attr_bitmask,
};
struct nfs_fattr fattr, dir_attr;
struct nfs4_link_res res = {
.server = server,
.fattr = &fattr,
.dir_attr = &dir_attr,
};
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LINK],
.rpc_argp = &arg,
.rpc_resp = &res,
};
int status;
nfs_fattr_init(res.fattr);
nfs_fattr_init(res.dir_attr);
status = rpc_call_sync(server->client, &msg, 0);
if (!status) {
update_changeattr(dir, &res.cinfo);
nfs_post_op_update_inode(dir, res.dir_attr);
nfs_post_op_update_inode(inode, res.fattr);
}
return status;
}
static int nfs4_proc_link(struct inode *inode, struct inode *dir, struct qstr *name)
{
struct nfs4_exception exception = { };
int err;
do {
err = nfs4_handle_exception(NFS_SERVER(inode),
_nfs4_proc_link(inode, dir, name),
&exception);
} while (exception.retry);
return err;
}
static int _nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
struct page *page, unsigned int len, struct iattr *sattr)
{
struct nfs_server *server = NFS_SERVER(dir);
struct nfs_fh fhandle;
struct nfs_fattr fattr, dir_fattr;
struct nfs4_create_arg arg = {
.dir_fh = NFS_FH(dir),
.server = server,
.name = &dentry->d_name,
.attrs = sattr,
.ftype = NF4LNK,
.bitmask = server->attr_bitmask,
};
struct nfs4_create_res res = {
.server = server,
.fh = &fhandle,
.fattr = &fattr,
.dir_fattr = &dir_fattr,
};
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SYMLINK],
.rpc_argp = &arg,
.rpc_resp = &res,
};
int status;
if (len > NFS4_MAXPATHLEN)
return -ENAMETOOLONG;
arg.u.symlink.pages = &page;
arg.u.symlink.len = len;
nfs_fattr_init(&fattr);
nfs_fattr_init(&dir_fattr);
status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
if (!status) {
update_changeattr(dir, &res.dir_cinfo);
nfs_post_op_update_inode(dir, res.dir_fattr);
status = nfs_instantiate(dentry, &fhandle, &fattr);
}
return status;
}
static int nfs4_proc_symlink(struct inode *dir, struct dentry *dentry,
struct page *page, unsigned int len, struct iattr *sattr)
{
struct nfs4_exception exception = { };
int err;
do {
err = nfs4_handle_exception(NFS_SERVER(dir),
_nfs4_proc_symlink(dir, dentry, page,
len, sattr),
&exception);
} while (exception.retry);
return err;
}
static int _nfs4_proc_mkdir(struct inode *dir, struct dentry *dentry,
struct iattr *sattr)
{
struct nfs_server *server = NFS_SERVER(dir);
struct nfs_fh fhandle;
struct nfs_fattr fattr, dir_fattr;
struct nfs4_create_arg arg = {
.dir_fh = NFS_FH(dir),
.server = server,
.name = &dentry->d_name,
.attrs = sattr,
.ftype = NF4DIR,
.bitmask = server->attr_bitmask,
};
struct nfs4_create_res res = {
.server = server,
.fh = &fhandle,
.fattr = &fattr,
.dir_fattr = &dir_fattr,
};
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CREATE],
.rpc_argp = &arg,
.rpc_resp = &res,
};
int status;
nfs_fattr_init(&fattr);
nfs_fattr_init(&dir_fattr);
status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
if (!status) {
update_changeattr(dir, &res.dir_cinfo);
nfs_post_op_update_inode(dir, res.dir_fattr);
status = nfs_instantiate(dentry, &fhandle, &fattr);
}
return status;
}
static int nfs4_proc_mkdir(struct inode *dir, struct dentry *dentry,
struct iattr *sattr)
{
struct nfs4_exception exception = { };
int err;
do {
err = nfs4_handle_exception(NFS_SERVER(dir),
_nfs4_proc_mkdir(dir, dentry, sattr),
&exception);
} while (exception.retry);
return err;
}
static int _nfs4_proc_readdir(struct dentry *dentry, struct rpc_cred *cred,
u64 cookie, struct page *page, unsigned int count, int plus)
{
struct inode *dir = dentry->d_inode;
struct nfs4_readdir_arg args = {
.fh = NFS_FH(dir),
.pages = &page,
.pgbase = 0,
.count = count,
.bitmask = NFS_SERVER(dentry->d_inode)->attr_bitmask,
};
struct nfs4_readdir_res res;
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READDIR],
.rpc_argp = &args,
.rpc_resp = &res,
.rpc_cred = cred,
};
int status;
dprintk("%s: dentry = %s/%s, cookie = %Lu\n", __func__,
dentry->d_parent->d_name.name,
dentry->d_name.name,
(unsigned long long)cookie);
nfs4_setup_readdir(cookie, NFS_COOKIEVERF(dir), dentry, &args);
res.pgbase = args.pgbase;
status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
if (status == 0)
memcpy(NFS_COOKIEVERF(dir), res.verifier.data, NFS4_VERIFIER_SIZE);
nfs_invalidate_atime(dir);
dprintk("%s: returns %d\n", __func__, status);
return status;
}
static int nfs4_proc_readdir(struct dentry *dentry, struct rpc_cred *cred,
u64 cookie, struct page *page, unsigned int count, int plus)
{
struct nfs4_exception exception = { };
int err;
do {
err = nfs4_handle_exception(NFS_SERVER(dentry->d_inode),
_nfs4_proc_readdir(dentry, cred, cookie,
page, count, plus),
&exception);
} while (exception.retry);
return err;
}
static int _nfs4_proc_mknod(struct inode *dir, struct dentry *dentry,
struct iattr *sattr, dev_t rdev)
{
struct nfs_server *server = NFS_SERVER(dir);
struct nfs_fh fh;
struct nfs_fattr fattr, dir_fattr;
struct nfs4_create_arg arg = {
.dir_fh = NFS_FH(dir),
.server = server,
.name = &dentry->d_name,
.attrs = sattr,
.bitmask = server->attr_bitmask,
};
struct nfs4_create_res res = {
.server = server,
.fh = &fh,
.fattr = &fattr,
.dir_fattr = &dir_fattr,
};
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CREATE],
.rpc_argp = &arg,
.rpc_resp = &res,
};
int status;
int mode = sattr->ia_mode;
nfs_fattr_init(&fattr);
nfs_fattr_init(&dir_fattr);
BUG_ON(!(sattr->ia_valid & ATTR_MODE));
BUG_ON(!S_ISFIFO(mode) && !S_ISBLK(mode) && !S_ISCHR(mode) && !S_ISSOCK(mode));
if (S_ISFIFO(mode))
arg.ftype = NF4FIFO;
else if (S_ISBLK(mode)) {
arg.ftype = NF4BLK;
arg.u.device.specdata1 = MAJOR(rdev);
arg.u.device.specdata2 = MINOR(rdev);
}
else if (S_ISCHR(mode)) {
arg.ftype = NF4CHR;
arg.u.device.specdata1 = MAJOR(rdev);
arg.u.device.specdata2 = MINOR(rdev);
}
else
arg.ftype = NF4SOCK;
status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
if (status == 0) {
update_changeattr(dir, &res.dir_cinfo);
nfs_post_op_update_inode(dir, res.dir_fattr);
status = nfs_instantiate(dentry, &fh, &fattr);
}
return status;
}
static int nfs4_proc_mknod(struct inode *dir, struct dentry *dentry,
struct iattr *sattr, dev_t rdev)
{
struct nfs4_exception exception = { };
int err;
do {
err = nfs4_handle_exception(NFS_SERVER(dir),
_nfs4_proc_mknod(dir, dentry, sattr, rdev),
&exception);
} while (exception.retry);
return err;
}
static int _nfs4_proc_statfs(struct nfs_server *server, struct nfs_fh *fhandle,
struct nfs_fsstat *fsstat)
{
struct nfs4_statfs_arg args = {
.fh = fhandle,
.bitmask = server->attr_bitmask,
};
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_STATFS],
.rpc_argp = &args,
.rpc_resp = fsstat,
};
nfs_fattr_init(fsstat->fattr);
return rpc_call_sync(server->client, &msg, 0);
}
static int nfs4_proc_statfs(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fsstat *fsstat)
{
struct nfs4_exception exception = { };
int err;
do {
err = nfs4_handle_exception(server,
_nfs4_proc_statfs(server, fhandle, fsstat),
&exception);
} while (exception.retry);
return err;
}
static int _nfs4_do_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle,
struct nfs_fsinfo *fsinfo)
{
struct nfs4_fsinfo_arg args = {
.fh = fhandle,
.bitmask = server->attr_bitmask,
};
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FSINFO],
.rpc_argp = &args,
.rpc_resp = fsinfo,
};
return rpc_call_sync(server->client, &msg, 0);
}
static int nfs4_do_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fsinfo *fsinfo)
{
struct nfs4_exception exception = { };
int err;
do {
err = nfs4_handle_exception(server,
_nfs4_do_fsinfo(server, fhandle, fsinfo),
&exception);
} while (exception.retry);
return err;
}
static int nfs4_proc_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle, struct nfs_fsinfo *fsinfo)
{
nfs_fattr_init(fsinfo->fattr);
return nfs4_do_fsinfo(server, fhandle, fsinfo);
}
static int _nfs4_proc_pathconf(struct nfs_server *server, struct nfs_fh *fhandle,
struct nfs_pathconf *pathconf)
{
struct nfs4_pathconf_arg args = {
.fh = fhandle,
.bitmask = server->attr_bitmask,
};
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_PATHCONF],
.rpc_argp = &args,
.rpc_resp = pathconf,
};
/* None of the pathconf attributes are mandatory to implement */
if ((args.bitmask[0] & nfs4_pathconf_bitmap[0]) == 0) {
memset(pathconf, 0, sizeof(*pathconf));
return 0;
}
nfs_fattr_init(pathconf->fattr);
return rpc_call_sync(server->client, &msg, 0);
}
static int nfs4_proc_pathconf(struct nfs_server *server, struct nfs_fh *fhandle,
struct nfs_pathconf *pathconf)
{
struct nfs4_exception exception = { };
int err;
do {
err = nfs4_handle_exception(server,
_nfs4_proc_pathconf(server, fhandle, pathconf),
&exception);
} while (exception.retry);
return err;
}
static int nfs4_read_done(struct rpc_task *task, struct nfs_read_data *data)
{
struct nfs_server *server = NFS_SERVER(data->inode);
if (nfs4_async_handle_error(task, server) == -EAGAIN) {
rpc_restart_call(task);
return -EAGAIN;
}
nfs_invalidate_atime(data->inode);
if (task->tk_status > 0)
renew_lease(server, data->timestamp);
return 0;
}
static void nfs4_proc_read_setup(struct nfs_read_data *data, struct rpc_message *msg)
{
data->timestamp = jiffies;
msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READ];
}
static int nfs4_write_done(struct rpc_task *task, struct nfs_write_data *data)
{
struct inode *inode = data->inode;
if (nfs4_async_handle_error(task, NFS_SERVER(inode)) == -EAGAIN) {
rpc_restart_call(task);
return -EAGAIN;
}
if (task->tk_status >= 0) {
renew_lease(NFS_SERVER(inode), data->timestamp);
nfs_post_op_update_inode_force_wcc(inode, data->res.fattr);
}
return 0;
}
static void nfs4_proc_write_setup(struct nfs_write_data *data, struct rpc_message *msg)
{
struct nfs_server *server = NFS_SERVER(data->inode);
data->args.bitmask = server->attr_bitmask;
data->res.server = server;
data->timestamp = jiffies;
msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_WRITE];
}
static int nfs4_commit_done(struct rpc_task *task, struct nfs_write_data *data)
{
struct inode *inode = data->inode;
if (nfs4_async_handle_error(task, NFS_SERVER(inode)) == -EAGAIN) {
rpc_restart_call(task);
return -EAGAIN;
}
nfs_refresh_inode(inode, data->res.fattr);
return 0;
}
static void nfs4_proc_commit_setup(struct nfs_write_data *data, struct rpc_message *msg)
{
struct nfs_server *server = NFS_SERVER(data->inode);
data->args.bitmask = server->attr_bitmask;
data->res.server = server;
msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_COMMIT];
}
/*
* nfs4_proc_async_renew(): This is not one of the nfs_rpc_ops; it is a special
* standalone procedure for queueing an asynchronous RENEW.
*/
static void nfs4_renew_done(struct rpc_task *task, void *data)
{
struct nfs_client *clp = (struct nfs_client *)task->tk_msg.rpc_argp;
unsigned long timestamp = (unsigned long)data;
if (task->tk_status < 0) {
switch (task->tk_status) {
case -NFS4ERR_STALE_CLIENTID:
case -NFS4ERR_EXPIRED:
case -NFS4ERR_CB_PATH_DOWN:
nfs4_schedule_state_recovery(clp);
}
return;
}
spin_lock(&clp->cl_lock);
if (time_before(clp->cl_last_renewal,timestamp))
clp->cl_last_renewal = timestamp;
spin_unlock(&clp->cl_lock);
}
static const struct rpc_call_ops nfs4_renew_ops = {
.rpc_call_done = nfs4_renew_done,
};
int nfs4_proc_async_renew(struct nfs_client *clp, struct rpc_cred *cred)
{
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RENEW],
.rpc_argp = clp,
.rpc_cred = cred,
};
return rpc_call_async(clp->cl_rpcclient, &msg, RPC_TASK_SOFT,
&nfs4_renew_ops, (void *)jiffies);
}
int nfs4_proc_renew(struct nfs_client *clp, struct rpc_cred *cred)
{
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RENEW],
.rpc_argp = clp,
.rpc_cred = cred,
};
unsigned long now = jiffies;
int status;
status = rpc_call_sync(clp->cl_rpcclient, &msg, 0);
if (status < 0)
return status;
spin_lock(&clp->cl_lock);
if (time_before(clp->cl_last_renewal,now))
clp->cl_last_renewal = now;
spin_unlock(&clp->cl_lock);
return 0;
}
static inline int nfs4_server_supports_acls(struct nfs_server *server)
{
return (server->caps & NFS_CAP_ACLS)
&& (server->acl_bitmask & ACL4_SUPPORT_ALLOW_ACL)
&& (server->acl_bitmask & ACL4_SUPPORT_DENY_ACL);
}
/* Assuming that XATTR_SIZE_MAX is a multiple of PAGE_CACHE_SIZE, and that
* it's OK to put sizeof(void) * (XATTR_SIZE_MAX/PAGE_CACHE_SIZE) bytes on
* the stack.
*/
#define NFS4ACL_MAXPAGES (XATTR_SIZE_MAX >> PAGE_CACHE_SHIFT)
static void buf_to_pages(const void *buf, size_t buflen,
struct page **pages, unsigned int *pgbase)
{
const void *p = buf;
*pgbase = offset_in_page(buf);
p -= *pgbase;
while (p < buf + buflen) {
*(pages++) = virt_to_page(p);
p += PAGE_CACHE_SIZE;
}
}
struct nfs4_cached_acl {
int cached;
size_t len;
char data[0];
};
static void nfs4_set_cached_acl(struct inode *inode, struct nfs4_cached_acl *acl)
{
struct nfs_inode *nfsi = NFS_I(inode);
spin_lock(&inode->i_lock);
kfree(nfsi->nfs4_acl);
nfsi->nfs4_acl = acl;
spin_unlock(&inode->i_lock);
}
static void nfs4_zap_acl_attr(struct inode *inode)
{
nfs4_set_cached_acl(inode, NULL);
}
static inline ssize_t nfs4_read_cached_acl(struct inode *inode, char *buf, size_t buflen)
{
struct nfs_inode *nfsi = NFS_I(inode);
struct nfs4_cached_acl *acl;
int ret = -ENOENT;
spin_lock(&inode->i_lock);
acl = nfsi->nfs4_acl;
if (acl == NULL)
goto out;
if (buf == NULL) /* user is just asking for length */
goto out_len;
if (acl->cached == 0)
goto out;
ret = -ERANGE; /* see getxattr(2) man page */
if (acl->len > buflen)
goto out;
memcpy(buf, acl->data, acl->len);
out_len:
ret = acl->len;
out:
spin_unlock(&inode->i_lock);
return ret;
}
static void nfs4_write_cached_acl(struct inode *inode, const char *buf, size_t acl_len)
{
struct nfs4_cached_acl *acl;
if (buf && acl_len <= PAGE_SIZE) {
acl = kmalloc(sizeof(*acl) + acl_len, GFP_KERNEL);
if (acl == NULL)
goto out;
acl->cached = 1;
memcpy(acl->data, buf, acl_len);
} else {
acl = kmalloc(sizeof(*acl), GFP_KERNEL);
if (acl == NULL)
goto out;
acl->cached = 0;
}
acl->len = acl_len;
out:
nfs4_set_cached_acl(inode, acl);
}
static ssize_t __nfs4_get_acl_uncached(struct inode *inode, void *buf, size_t buflen)
{
struct page *pages[NFS4ACL_MAXPAGES];
struct nfs_getaclargs args = {
.fh = NFS_FH(inode),
.acl_pages = pages,
.acl_len = buflen,
};
size_t resp_len = buflen;
void *resp_buf;
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_GETACL],
.rpc_argp = &args,
.rpc_resp = &resp_len,
};
struct page *localpage = NULL;
int ret;
if (buflen < PAGE_SIZE) {
/* As long as we're doing a round trip to the server anyway,
* let's be prepared for a page of acl data. */
localpage = alloc_page(GFP_KERNEL);
resp_buf = page_address(localpage);
if (localpage == NULL)
return -ENOMEM;
args.acl_pages[0] = localpage;
args.acl_pgbase = 0;
resp_len = args.acl_len = PAGE_SIZE;
} else {
resp_buf = buf;
buf_to_pages(buf, buflen, args.acl_pages, &args.acl_pgbase);
}
ret = rpc_call_sync(NFS_CLIENT(inode), &msg, 0);
if (ret)
goto out_free;
if (resp_len > args.acl_len)
nfs4_write_cached_acl(inode, NULL, resp_len);
else
nfs4_write_cached_acl(inode, resp_buf, resp_len);
if (buf) {
ret = -ERANGE;
if (resp_len > buflen)
goto out_free;
if (localpage)
memcpy(buf, resp_buf, resp_len);
}
ret = resp_len;
out_free:
if (localpage)
__free_page(localpage);
return ret;
}
static ssize_t nfs4_get_acl_uncached(struct inode *inode, void *buf, size_t buflen)
{
struct nfs4_exception exception = { };
ssize_t ret;
do {
ret = __nfs4_get_acl_uncached(inode, buf, buflen);
if (ret >= 0)
break;
ret = nfs4_handle_exception(NFS_SERVER(inode), ret, &exception);
} while (exception.retry);
return ret;
}
static ssize_t nfs4_proc_get_acl(struct inode *inode, void *buf, size_t buflen)
{
struct nfs_server *server = NFS_SERVER(inode);
int ret;
if (!nfs4_server_supports_acls(server))
return -EOPNOTSUPP;
ret = nfs_revalidate_inode(server, inode);
if (ret < 0)
return ret;
ret = nfs4_read_cached_acl(inode, buf, buflen);
if (ret != -ENOENT)
return ret;
return nfs4_get_acl_uncached(inode, buf, buflen);
}
static int __nfs4_proc_set_acl(struct inode *inode, const void *buf, size_t buflen)
{
struct nfs_server *server = NFS_SERVER(inode);
struct page *pages[NFS4ACL_MAXPAGES];
struct nfs_setaclargs arg = {
.fh = NFS_FH(inode),
.acl_pages = pages,
.acl_len = buflen,
};
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETACL],
.rpc_argp = &arg,
.rpc_resp = NULL,
};
int ret;
if (!nfs4_server_supports_acls(server))
return -EOPNOTSUPP;
nfs_inode_return_delegation(inode);
buf_to_pages(buf, buflen, arg.acl_pages, &arg.acl_pgbase);
ret = rpc_call_sync(NFS_CLIENT(inode), &msg, 0);
nfs_zap_caches(inode);
return ret;
}
static int nfs4_proc_set_acl(struct inode *inode, const void *buf, size_t buflen)
{
struct nfs4_exception exception = { };
int err;
do {
err = nfs4_handle_exception(NFS_SERVER(inode),
__nfs4_proc_set_acl(inode, buf, buflen),
&exception);
} while (exception.retry);
return err;
}
static int
nfs4_async_handle_error(struct rpc_task *task, const struct nfs_server *server)
{
struct nfs_client *clp = server->nfs_client;
if (!clp || task->tk_status >= 0)
return 0;
switch(task->tk_status) {
case -NFS4ERR_STALE_CLIENTID:
case -NFS4ERR_STALE_STATEID:
case -NFS4ERR_EXPIRED:
rpc_sleep_on(&clp->cl_rpcwaitq, task, NULL);
nfs4_schedule_state_recovery(clp);
if (test_bit(NFS4CLNT_STATE_RECOVER, &clp->cl_state) == 0)
rpc_wake_up_queued_task(&clp->cl_rpcwaitq, task);
task->tk_status = 0;
return -EAGAIN;
case -NFS4ERR_DELAY:
nfs_inc_server_stats((struct nfs_server *) server,
NFSIOS_DELAY);
case -NFS4ERR_GRACE:
rpc_delay(task, NFS4_POLL_RETRY_MAX);
task->tk_status = 0;
return -EAGAIN;
case -NFS4ERR_OLD_STATEID:
task->tk_status = 0;
return -EAGAIN;
}
task->tk_status = nfs4_map_errors(task->tk_status);
return 0;
}
static int nfs4_wait_bit_killable(void *word)
{
if (fatal_signal_pending(current))
return -ERESTARTSYS;
schedule();
return 0;
}
static int nfs4_wait_clnt_recover(struct rpc_clnt *clnt, struct nfs_client *clp)
{
int res;
might_sleep();
rwsem_acquire(&clp->cl_sem.dep_map, 0, 0, _RET_IP_);
res = wait_on_bit(&clp->cl_state, NFS4CLNT_STATE_RECOVER,
nfs4_wait_bit_killable, TASK_KILLABLE);
rwsem_release(&clp->cl_sem.dep_map, 1, _RET_IP_);
return res;
}
static int nfs4_delay(struct rpc_clnt *clnt, long *timeout)
{
int res = 0;
might_sleep();
if (*timeout <= 0)
*timeout = NFS4_POLL_RETRY_MIN;
if (*timeout > NFS4_POLL_RETRY_MAX)
*timeout = NFS4_POLL_RETRY_MAX;
schedule_timeout_killable(*timeout);
if (fatal_signal_pending(current))
res = -ERESTARTSYS;
*timeout <<= 1;
return res;
}
/* This is the error handling routine for processes that are allowed
* to sleep.
*/
static int nfs4_handle_exception(const struct nfs_server *server, int errorcode, struct nfs4_exception *exception)
{
struct nfs_client *clp = server->nfs_client;
int ret = errorcode;
exception->retry = 0;
switch(errorcode) {
case 0:
return 0;
case -NFS4ERR_STALE_CLIENTID:
case -NFS4ERR_STALE_STATEID:
case -NFS4ERR_EXPIRED:
nfs4_schedule_state_recovery(clp);
ret = nfs4_wait_clnt_recover(server->client, clp);
if (ret == 0)
exception->retry = 1;
break;
case -NFS4ERR_FILE_OPEN:
case -NFS4ERR_GRACE:
case -NFS4ERR_DELAY:
ret = nfs4_delay(server->client, &exception->timeout);
if (ret != 0)
break;
case -NFS4ERR_OLD_STATEID:
exception->retry = 1;
}
/* We failed to handle the error */
return nfs4_map_errors(ret);
}
int nfs4_proc_setclientid(struct nfs_client *clp, u32 program, unsigned short port, struct rpc_cred *cred)
{
nfs4_verifier sc_verifier;
struct nfs4_setclientid setclientid = {
.sc_verifier = &sc_verifier,
.sc_prog = program,
};
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETCLIENTID],
.rpc_argp = &setclientid,
.rpc_resp = clp,
.rpc_cred = cred,
};
__be32 *p;
int loop = 0;
int status;
p = (__be32*)sc_verifier.data;
*p++ = htonl((u32)clp->cl_boot_time.tv_sec);
*p = htonl((u32)clp->cl_boot_time.tv_nsec);
for(;;) {
setclientid.sc_name_len = scnprintf(setclientid.sc_name,
sizeof(setclientid.sc_name), "%s/%s %s %s %u",
clp->cl_ipaddr,
rpc_peeraddr2str(clp->cl_rpcclient,
RPC_DISPLAY_ADDR),
rpc_peeraddr2str(clp->cl_rpcclient,
RPC_DISPLAY_PROTO),
clp->cl_rpcclient->cl_auth->au_ops->au_name,
clp->cl_id_uniquifier);
setclientid.sc_netid_len = scnprintf(setclientid.sc_netid,
sizeof(setclientid.sc_netid),
rpc_peeraddr2str(clp->cl_rpcclient,
RPC_DISPLAY_NETID));
setclientid.sc_uaddr_len = scnprintf(setclientid.sc_uaddr,
sizeof(setclientid.sc_uaddr), "%s.%u.%u",
clp->cl_ipaddr, port >> 8, port & 255);
status = rpc_call_sync(clp->cl_rpcclient, &msg, 0);
if (status != -NFS4ERR_CLID_INUSE)
break;
if (signalled())
break;
if (loop++ & 1)
ssleep(clp->cl_lease_time + 1);
else
if (++clp->cl_id_uniquifier == 0)
break;
}
return status;
}
static int _nfs4_proc_setclientid_confirm(struct nfs_client *clp, struct rpc_cred *cred)
{
struct nfs_fsinfo fsinfo;
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETCLIENTID_CONFIRM],
.rpc_argp = clp,
.rpc_resp = &fsinfo,
.rpc_cred = cred,
};
unsigned long now;
int status;
now = jiffies;
status = rpc_call_sync(clp->cl_rpcclient, &msg, 0);
if (status == 0) {
spin_lock(&clp->cl_lock);
clp->cl_lease_time = fsinfo.lease_time * HZ;
clp->cl_last_renewal = now;
clear_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
spin_unlock(&clp->cl_lock);
}
return status;
}
int nfs4_proc_setclientid_confirm(struct nfs_client *clp, struct rpc_cred *cred)
{
long timeout;
int err;
do {
err = _nfs4_proc_setclientid_confirm(clp, cred);
switch (err) {
case 0:
return err;
case -NFS4ERR_RESOURCE:
/* The IBM lawyers misread another document! */
case -NFS4ERR_DELAY:
err = nfs4_delay(clp->cl_rpcclient, &timeout);
}
} while (err == 0);
return err;
}
struct nfs4_delegreturndata {
struct nfs4_delegreturnargs args;
struct nfs4_delegreturnres res;
struct nfs_fh fh;
nfs4_stateid stateid;
unsigned long timestamp;
struct nfs_fattr fattr;
int rpc_status;
};
static void nfs4_delegreturn_done(struct rpc_task *task, void *calldata)
{
struct nfs4_delegreturndata *data = calldata;
data->rpc_status = task->tk_status;
if (data->rpc_status == 0)
renew_lease(data->res.server, data->timestamp);
}
static void nfs4_delegreturn_release(void *calldata)
{
kfree(calldata);
}
static const struct rpc_call_ops nfs4_delegreturn_ops = {
.rpc_call_done = nfs4_delegreturn_done,
.rpc_release = nfs4_delegreturn_release,
};
static int _nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid, int issync)
{
struct nfs4_delegreturndata *data;
struct nfs_server *server = NFS_SERVER(inode);
struct rpc_task *task;
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_DELEGRETURN],
.rpc_cred = cred,
};
struct rpc_task_setup task_setup_data = {
.rpc_client = server->client,
.rpc_message = &msg,
.callback_ops = &nfs4_delegreturn_ops,
.flags = RPC_TASK_ASYNC,
};
int status = 0;
data = kmalloc(sizeof(*data), GFP_KERNEL);
if (data == NULL)
return -ENOMEM;
data->args.fhandle = &data->fh;
data->args.stateid = &data->stateid;
data->args.bitmask = server->attr_bitmask;
nfs_copy_fh(&data->fh, NFS_FH(inode));
memcpy(&data->stateid, stateid, sizeof(data->stateid));
data->res.fattr = &data->fattr;
data->res.server = server;
nfs_fattr_init(data->res.fattr);
data->timestamp = jiffies;
data->rpc_status = 0;
task_setup_data.callback_data = data;
msg.rpc_argp = &data->args,
msg.rpc_resp = &data->res,
task = rpc_run_task(&task_setup_data);
if (IS_ERR(task))
return PTR_ERR(task);
if (!issync)
goto out;
status = nfs4_wait_for_completion_rpc_task(task);
if (status != 0)
goto out;
status = data->rpc_status;
if (status != 0)
goto out;
nfs_refresh_inode(inode, &data->fattr);
out:
rpc_put_task(task);
return status;
}
int nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid, int issync)
{
struct nfs_server *server = NFS_SERVER(inode);
struct nfs4_exception exception = { };
int err;
do {
err = _nfs4_proc_delegreturn(inode, cred, stateid, issync);
switch (err) {
case -NFS4ERR_STALE_STATEID:
case -NFS4ERR_EXPIRED:
case 0:
return 0;
}
err = nfs4_handle_exception(server, err, &exception);
} while (exception.retry);
return err;
}
#define NFS4_LOCK_MINTIMEOUT (1 * HZ)
#define NFS4_LOCK_MAXTIMEOUT (30 * HZ)
/*
* sleep, with exponential backoff, and retry the LOCK operation.
*/
static unsigned long
nfs4_set_lock_task_retry(unsigned long timeout)
{
schedule_timeout_killable(timeout);
timeout <<= 1;
if (timeout > NFS4_LOCK_MAXTIMEOUT)
return NFS4_LOCK_MAXTIMEOUT;
return timeout;
}
static int _nfs4_proc_getlk(struct nfs4_state *state, int cmd, struct file_lock *request)
{
struct inode *inode = state->inode;
struct nfs_server *server = NFS_SERVER(inode);
struct nfs_client *clp = server->nfs_client;
struct nfs_lockt_args arg = {
.fh = NFS_FH(inode),
.fl = request,
};
struct nfs_lockt_res res = {
.denied = request,
};
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOCKT],
.rpc_argp = &arg,
.rpc_resp = &res,
.rpc_cred = state->owner->so_cred,
};
struct nfs4_lock_state *lsp;
int status;
down_read(&clp->cl_sem);
arg.lock_owner.clientid = clp->cl_clientid;
status = nfs4_set_lock_state(state, request);
if (status != 0)
goto out;
lsp = request->fl_u.nfs4_fl.owner;
arg.lock_owner.id = lsp->ls_id.id;
status = rpc_call_sync(server->client, &msg, 0);
switch (status) {
case 0:
request->fl_type = F_UNLCK;
break;
case -NFS4ERR_DENIED:
status = 0;
}
request->fl_ops->fl_release_private(request);
out:
up_read(&clp->cl_sem);
return status;
}
static int nfs4_proc_getlk(struct nfs4_state *state, int cmd, struct file_lock *request)
{
struct nfs4_exception exception = { };
int err;
do {
err = nfs4_handle_exception(NFS_SERVER(state->inode),
_nfs4_proc_getlk(state, cmd, request),
&exception);
} while (exception.retry);
return err;
}
static int do_vfs_lock(struct file *file, struct file_lock *fl)
{
int res = 0;
switch (fl->fl_flags & (FL_POSIX|FL_FLOCK)) {
case FL_POSIX:
res = posix_lock_file_wait(file, fl);
break;
case FL_FLOCK:
res = flock_lock_file_wait(file, fl);
break;
default:
BUG();
}
return res;
}
struct nfs4_unlockdata {
struct nfs_locku_args arg;
struct nfs_locku_res res;
struct nfs4_lock_state *lsp;
struct nfs_open_context *ctx;
struct file_lock fl;
const struct nfs_server *server;
unsigned long timestamp;
};
static struct nfs4_unlockdata *nfs4_alloc_unlockdata(struct file_lock *fl,
struct nfs_open_context *ctx,
struct nfs4_lock_state *lsp,
struct nfs_seqid *seqid)
{
struct nfs4_unlockdata *p;
struct inode *inode = lsp->ls_state->inode;
p = kmalloc(sizeof(*p), GFP_KERNEL);
if (p == NULL)
return NULL;
p->arg.fh = NFS_FH(inode);
p->arg.fl = &p->fl;
p->arg.seqid = seqid;
p->res.seqid = seqid;
p->arg.stateid = &lsp->ls_stateid;
p->lsp = lsp;
atomic_inc(&lsp->ls_count);
/* Ensure we don't close file until we're done freeing locks! */
p->ctx = get_nfs_open_context(ctx);
memcpy(&p->fl, fl, sizeof(p->fl));
p->server = NFS_SERVER(inode);
return p;
}
static void nfs4_locku_release_calldata(void *data)
{
struct nfs4_unlockdata *calldata = data;
nfs_free_seqid(calldata->arg.seqid);
nfs4_put_lock_state(calldata->lsp);
put_nfs_open_context(calldata->ctx);
kfree(calldata);
}
static void nfs4_locku_done(struct rpc_task *task, void *data)
{
struct nfs4_unlockdata *calldata = data;
if (RPC_ASSASSINATED(task))
return;
switch (task->tk_status) {
case 0:
memcpy(calldata->lsp->ls_stateid.data,
calldata->res.stateid.data,
sizeof(calldata->lsp->ls_stateid.data));
renew_lease(calldata->server, calldata->timestamp);
break;
case -NFS4ERR_STALE_STATEID:
case -NFS4ERR_EXPIRED:
break;
default:
if (nfs4_async_handle_error(task, calldata->server) == -EAGAIN)
rpc_restart_call(task);
}
}
static void nfs4_locku_prepare(struct rpc_task *task, void *data)
{
struct nfs4_unlockdata *calldata = data;
if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
return;
if ((calldata->lsp->ls_flags & NFS_LOCK_INITIALIZED) == 0) {
/* Note: exit _without_ running nfs4_locku_done */
task->tk_action = NULL;
return;
}
calldata->timestamp = jiffies;
rpc_call_start(task);
}
static const struct rpc_call_ops nfs4_locku_ops = {
.rpc_call_prepare = nfs4_locku_prepare,
.rpc_call_done = nfs4_locku_done,
.rpc_release = nfs4_locku_release_calldata,
};
static struct rpc_task *nfs4_do_unlck(struct file_lock *fl,
struct nfs_open_context *ctx,
struct nfs4_lock_state *lsp,
struct nfs_seqid *seqid)
{
struct nfs4_unlockdata *data;
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOCKU],
.rpc_cred = ctx->cred,
};
struct rpc_task_setup task_setup_data = {
.rpc_client = NFS_CLIENT(lsp->ls_state->inode),
.rpc_message = &msg,
.callback_ops = &nfs4_locku_ops,
.workqueue = nfsiod_workqueue,
.flags = RPC_TASK_ASYNC,
};
NFSv4: Make sure unlock is really an unlock when cancelling a lock I ran into a curious issue when a lock is being canceled. The cancellation results in a lock request to the vfs layer instead of an unlock request. This is particularly insidious when the process that owns the lock is exiting. In that case, sometimes the erroneous lock is applied AFTER the process has entered zombie state, preventing the lock from ever being released. Eventually other processes block on the lock causing a slow degredation of the system. In the 2.6.16 kernel this was investigated on, the problem is compounded by the fact that the cl_sem is held while blocking on the vfs lock, which results in most processes accessing the nfs file system in question hanging. In more detail, here is how the situation occurs: first _nfs4_do_setlk(): static int _nfs4_do_setlk(struct nfs4_state *state, int cmd, struct file_lock *fl, int reclaim) ... ret = nfs4_wait_for_completion_rpc_task(task); if (ret == 0) { ... } else data->cancelled = 1; then nfs4_lock_release(): static void nfs4_lock_release(void *calldata) ... if (data->cancelled != 0) { struct rpc_task *task; task = nfs4_do_unlck(&data->fl, data->ctx, data->lsp, data->arg.lock_seqid); The problem is the same file_lock that was passed in to _nfs4_do_setlk() gets passed to nfs4_do_unlck() from nfs4_lock_release(). So the type is still F_RDLCK or FWRLCK, not F_UNLCK. At some point, when cancelling the lock, the type needs to be changed to F_UNLCK. It seemed easiest to do that in nfs4_do_unlck(), but it could be done in nfs4_lock_release(). The concern I had with doing it there was if something still needed the original file_lock, though it turns out the original file_lock still needs to be modified by nfs4_do_unlck() because nfs4_do_unlck() uses the original file_lock to pass to the vfs layer, and a copy of the original file_lock for the RPC request. It seems like the simplest solution is to force all situations where nfs4_do_unlck() is being used to result in an unlock, so with that in mind, I made the following change: Signed-off-by: Frank Filz <ffilzlnx@us.ibm.com> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
2007-07-09 22:32:29 +00:00
/* Ensure this is an unlock - when canceling a lock, the
* canceled lock is passed in, and it won't be an unlock.
*/
fl->fl_type = F_UNLCK;
data = nfs4_alloc_unlockdata(fl, ctx, lsp, seqid);
if (data == NULL) {
nfs_free_seqid(seqid);
return ERR_PTR(-ENOMEM);
}
msg.rpc_argp = &data->arg,
msg.rpc_resp = &data->res,
task_setup_data.callback_data = data;
return rpc_run_task(&task_setup_data);
}
static int nfs4_proc_unlck(struct nfs4_state *state, int cmd, struct file_lock *request)
{
struct nfs_seqid *seqid;
struct nfs4_lock_state *lsp;
struct rpc_task *task;
int status = 0;
unsigned char fl_flags = request->fl_flags;
status = nfs4_set_lock_state(state, request);
/* Unlock _before_ we do the RPC call */
request->fl_flags |= FL_EXISTS;
if (do_vfs_lock(request->fl_file, request) == -ENOENT)
goto out;
if (status != 0)
goto out;
/* Is this a delegated lock? */
if (test_bit(NFS_DELEGATED_STATE, &state->flags))
goto out;
lsp = request->fl_u.nfs4_fl.owner;
seqid = nfs_alloc_seqid(&lsp->ls_seqid);
status = -ENOMEM;
if (seqid == NULL)
goto out;
task = nfs4_do_unlck(request, nfs_file_open_context(request->fl_file), lsp, seqid);
status = PTR_ERR(task);
if (IS_ERR(task))
goto out;
status = nfs4_wait_for_completion_rpc_task(task);
rpc_put_task(task);
out:
request->fl_flags = fl_flags;
return status;
}
struct nfs4_lockdata {
struct nfs_lock_args arg;
struct nfs_lock_res res;
struct nfs4_lock_state *lsp;
struct nfs_open_context *ctx;
struct file_lock fl;
unsigned long timestamp;
int rpc_status;
int cancelled;
};
static struct nfs4_lockdata *nfs4_alloc_lockdata(struct file_lock *fl,
struct nfs_open_context *ctx, struct nfs4_lock_state *lsp)
{
struct nfs4_lockdata *p;
struct inode *inode = lsp->ls_state->inode;
struct nfs_server *server = NFS_SERVER(inode);
p = kzalloc(sizeof(*p), GFP_KERNEL);
if (p == NULL)
return NULL;
p->arg.fh = NFS_FH(inode);
p->arg.fl = &p->fl;
p->arg.open_seqid = nfs_alloc_seqid(&lsp->ls_state->owner->so_seqid);
if (p->arg.open_seqid == NULL)
goto out_free;
p->arg.lock_seqid = nfs_alloc_seqid(&lsp->ls_seqid);
if (p->arg.lock_seqid == NULL)
goto out_free_seqid;
p->arg.lock_stateid = &lsp->ls_stateid;
p->arg.lock_owner.clientid = server->nfs_client->cl_clientid;
p->arg.lock_owner.id = lsp->ls_id.id;
p->res.lock_seqid = p->arg.lock_seqid;
p->lsp = lsp;
atomic_inc(&lsp->ls_count);
p->ctx = get_nfs_open_context(ctx);
memcpy(&p->fl, fl, sizeof(p->fl));
return p;
out_free_seqid:
nfs_free_seqid(p->arg.open_seqid);
out_free:
kfree(p);
return NULL;
}
static void nfs4_lock_prepare(struct rpc_task *task, void *calldata)
{
struct nfs4_lockdata *data = calldata;
struct nfs4_state *state = data->lsp->ls_state;
dprintk("%s: begin!\n", __func__);
if (nfs_wait_on_sequence(data->arg.lock_seqid, task) != 0)
return;
/* Do we need to do an open_to_lock_owner? */
if (!(data->arg.lock_seqid->sequence->flags & NFS_SEQID_CONFIRMED)) {
if (nfs_wait_on_sequence(data->arg.open_seqid, task) != 0)
return;
data->arg.open_stateid = &state->stateid;
data->arg.new_lock_owner = 1;
data->res.open_seqid = data->arg.open_seqid;
} else
data->arg.new_lock_owner = 0;
data->timestamp = jiffies;
rpc_call_start(task);
dprintk("%s: done!, ret = %d\n", __func__, data->rpc_status);
}
static void nfs4_lock_done(struct rpc_task *task, void *calldata)
{
struct nfs4_lockdata *data = calldata;
dprintk("%s: begin!\n", __func__);
data->rpc_status = task->tk_status;
if (RPC_ASSASSINATED(task))
goto out;
if (data->arg.new_lock_owner != 0) {
if (data->rpc_status == 0)
nfs_confirm_seqid(&data->lsp->ls_seqid, 0);
else
goto out;
}
if (data->rpc_status == 0) {
memcpy(data->lsp->ls_stateid.data, data->res.stateid.data,
sizeof(data->lsp->ls_stateid.data));
data->lsp->ls_flags |= NFS_LOCK_INITIALIZED;
renew_lease(NFS_SERVER(data->ctx->path.dentry->d_inode), data->timestamp);
}
out:
dprintk("%s: done, ret = %d!\n", __func__, data->rpc_status);
}
static void nfs4_lock_release(void *calldata)
{
struct nfs4_lockdata *data = calldata;
dprintk("%s: begin!\n", __func__);
nfs_free_seqid(data->arg.open_seqid);
if (data->cancelled != 0) {
struct rpc_task *task;
task = nfs4_do_unlck(&data->fl, data->ctx, data->lsp,
data->arg.lock_seqid);
if (!IS_ERR(task))
rpc_put_task(task);
dprintk("%s: cancelling lock!\n", __func__);
} else
nfs_free_seqid(data->arg.lock_seqid);
nfs4_put_lock_state(data->lsp);
put_nfs_open_context(data->ctx);
kfree(data);
dprintk("%s: done!\n", __func__);
}
static const struct rpc_call_ops nfs4_lock_ops = {
.rpc_call_prepare = nfs4_lock_prepare,
.rpc_call_done = nfs4_lock_done,
.rpc_release = nfs4_lock_release,
};
static int _nfs4_do_setlk(struct nfs4_state *state, int cmd, struct file_lock *fl, int reclaim)
{
struct nfs4_lockdata *data;
struct rpc_task *task;
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOCK],
.rpc_cred = state->owner->so_cred,
};
struct rpc_task_setup task_setup_data = {
.rpc_client = NFS_CLIENT(state->inode),
.rpc_message = &msg,
.callback_ops = &nfs4_lock_ops,
.workqueue = nfsiod_workqueue,
.flags = RPC_TASK_ASYNC,
};
int ret;
dprintk("%s: begin!\n", __func__);
data = nfs4_alloc_lockdata(fl, nfs_file_open_context(fl->fl_file),
fl->fl_u.nfs4_fl.owner);
if (data == NULL)
return -ENOMEM;
if (IS_SETLKW(cmd))
data->arg.block = 1;
if (reclaim != 0)
data->arg.reclaim = 1;
msg.rpc_argp = &data->arg,
msg.rpc_resp = &data->res,
task_setup_data.callback_data = data;
task = rpc_run_task(&task_setup_data);
if (IS_ERR(task))
return PTR_ERR(task);
ret = nfs4_wait_for_completion_rpc_task(task);
if (ret == 0) {
ret = data->rpc_status;
if (ret == -NFS4ERR_DENIED)
ret = -EAGAIN;
} else
data->cancelled = 1;
rpc_put_task(task);
dprintk("%s: done, ret = %d!\n", __func__, ret);
return ret;
}
static int nfs4_lock_reclaim(struct nfs4_state *state, struct file_lock *request)
{
struct nfs_server *server = NFS_SERVER(state->inode);
struct nfs4_exception exception = { };
int err;
do {
/* Cache the lock if possible... */
if (test_bit(NFS_DELEGATED_STATE, &state->flags) != 0)
return 0;
err = _nfs4_do_setlk(state, F_SETLK, request, 1);
if (err != -NFS4ERR_DELAY)
break;
nfs4_handle_exception(server, err, &exception);
} while (exception.retry);
return err;
}
static int nfs4_lock_expired(struct nfs4_state *state, struct file_lock *request)
{
struct nfs_server *server = NFS_SERVER(state->inode);
struct nfs4_exception exception = { };
int err;
err = nfs4_set_lock_state(state, request);
if (err != 0)
return err;
do {
if (test_bit(NFS_DELEGATED_STATE, &state->flags) != 0)
return 0;
err = _nfs4_do_setlk(state, F_SETLK, request, 0);
if (err != -NFS4ERR_DELAY)
break;
nfs4_handle_exception(server, err, &exception);
} while (exception.retry);
return err;
}
static int _nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
{
struct nfs_client *clp = state->owner->so_client;
unsigned char fl_flags = request->fl_flags;
int status;
/* Is this a delegated open? */
status = nfs4_set_lock_state(state, request);
if (status != 0)
goto out;
request->fl_flags |= FL_ACCESS;
status = do_vfs_lock(request->fl_file, request);
if (status < 0)
goto out;
down_read(&clp->cl_sem);
if (test_bit(NFS_DELEGATED_STATE, &state->flags)) {
struct nfs_inode *nfsi = NFS_I(state->inode);
/* Yes: cache locks! */
down_read(&nfsi->rwsem);
/* ...but avoid races with delegation recall... */
if (test_bit(NFS_DELEGATED_STATE, &state->flags)) {
request->fl_flags = fl_flags & ~FL_SLEEP;
status = do_vfs_lock(request->fl_file, request);
up_read(&nfsi->rwsem);
goto out_unlock;
}
up_read(&nfsi->rwsem);
}
status = _nfs4_do_setlk(state, cmd, request, 0);
if (status != 0)
goto out_unlock;
/* Note: we always want to sleep here! */
request->fl_flags = fl_flags | FL_SLEEP;
if (do_vfs_lock(request->fl_file, request) < 0)
printk(KERN_WARNING "%s: VFS is out of sync with lock manager!\n", __func__);
out_unlock:
up_read(&clp->cl_sem);
out:
request->fl_flags = fl_flags;
return status;
}
static int nfs4_proc_setlk(struct nfs4_state *state, int cmd, struct file_lock *request)
{
struct nfs4_exception exception = { };
int err;
do {
err = nfs4_handle_exception(NFS_SERVER(state->inode),
_nfs4_proc_setlk(state, cmd, request),
&exception);
} while (exception.retry);
return err;
}
static int
nfs4_proc_lock(struct file *filp, int cmd, struct file_lock *request)
{
struct nfs_open_context *ctx;
struct nfs4_state *state;
unsigned long timeout = NFS4_LOCK_MINTIMEOUT;
int status;
/* verify open state */
ctx = nfs_file_open_context(filp);
state = ctx->state;
if (request->fl_start < 0 || request->fl_end < 0)
return -EINVAL;
if (IS_GETLK(cmd))
return nfs4_proc_getlk(state, F_GETLK, request);
if (!(IS_SETLK(cmd) || IS_SETLKW(cmd)))
return -EINVAL;
if (request->fl_type == F_UNLCK)
return nfs4_proc_unlck(state, cmd, request);
do {
status = nfs4_proc_setlk(state, cmd, request);
if ((status != -EAGAIN) || IS_SETLK(cmd))
break;
timeout = nfs4_set_lock_task_retry(timeout);
status = -ERESTARTSYS;
if (signalled())
break;
} while(status < 0);
return status;
}
int nfs4_lock_delegation_recall(struct nfs4_state *state, struct file_lock *fl)
{
struct nfs_server *server = NFS_SERVER(state->inode);
struct nfs4_exception exception = { };
int err;
err = nfs4_set_lock_state(state, fl);
if (err != 0)
goto out;
do {
err = _nfs4_do_setlk(state, F_SETLK, fl, 0);
if (err != -NFS4ERR_DELAY)
break;
err = nfs4_handle_exception(server, err, &exception);
} while (exception.retry);
out:
return err;
}
#define XATTR_NAME_NFSV4_ACL "system.nfs4_acl"
int nfs4_setxattr(struct dentry *dentry, const char *key, const void *buf,
size_t buflen, int flags)
{
struct inode *inode = dentry->d_inode;
if (strcmp(key, XATTR_NAME_NFSV4_ACL) != 0)
return -EOPNOTSUPP;
return nfs4_proc_set_acl(inode, buf, buflen);
}
/* The getxattr man page suggests returning -ENODATA for unknown attributes,
* and that's what we'll do for e.g. user attributes that haven't been set.
* But we'll follow ext2/ext3's lead by returning -EOPNOTSUPP for unsupported
* attributes in kernel-managed attribute namespaces. */
ssize_t nfs4_getxattr(struct dentry *dentry, const char *key, void *buf,
size_t buflen)
{
struct inode *inode = dentry->d_inode;
if (strcmp(key, XATTR_NAME_NFSV4_ACL) != 0)
return -EOPNOTSUPP;
return nfs4_proc_get_acl(inode, buf, buflen);
}
ssize_t nfs4_listxattr(struct dentry *dentry, char *buf, size_t buflen)
{
size_t len = strlen(XATTR_NAME_NFSV4_ACL) + 1;
if (!nfs4_server_supports_acls(NFS_SERVER(dentry->d_inode)))
return 0;
if (buf && buflen < len)
return -ERANGE;
if (buf)
memcpy(buf, XATTR_NAME_NFSV4_ACL, len);
return len;
}
int nfs4_proc_fs_locations(struct inode *dir, const struct qstr *name,
struct nfs4_fs_locations *fs_locations, struct page *page)
{
struct nfs_server *server = NFS_SERVER(dir);
u32 bitmask[2] = {
[0] = FATTR4_WORD0_FSID | FATTR4_WORD0_FS_LOCATIONS,
[1] = FATTR4_WORD1_MOUNTED_ON_FILEID,
};
struct nfs4_fs_locations_arg args = {
.dir_fh = NFS_FH(dir),
.name = name,
.page = page,
.bitmask = bitmask,
};
struct rpc_message msg = {
.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_FS_LOCATIONS],
.rpc_argp = &args,
.rpc_resp = fs_locations,
};
int status;
dprintk("%s: start\n", __func__);
nfs_fattr_init(&fs_locations->fattr);
fs_locations->server = server;
fs_locations->nlocations = 0;
status = rpc_call_sync(server->client, &msg, 0);
dprintk("%s: returned status = %d\n", __func__, status);
return status;
}
struct nfs4_state_recovery_ops nfs4_reboot_recovery_ops = {
.recover_open = nfs4_open_reclaim,
.recover_lock = nfs4_lock_reclaim,
};
struct nfs4_state_recovery_ops nfs4_network_partition_recovery_ops = {
.recover_open = nfs4_open_expired,
.recover_lock = nfs4_lock_expired,
};
static const struct inode_operations nfs4_file_inode_operations = {
.permission = nfs_permission,
.getattr = nfs_getattr,
.setattr = nfs_setattr,
.getxattr = nfs4_getxattr,
.setxattr = nfs4_setxattr,
.listxattr = nfs4_listxattr,
};
const struct nfs_rpc_ops nfs_v4_clientops = {
.version = 4, /* protocol version */
.dentry_ops = &nfs4_dentry_operations,
.dir_inode_ops = &nfs4_dir_inode_operations,
.file_inode_ops = &nfs4_file_inode_operations,
.getroot = nfs4_proc_get_root,
.getattr = nfs4_proc_getattr,
.setattr = nfs4_proc_setattr,
.lookupfh = nfs4_proc_lookupfh,
.lookup = nfs4_proc_lookup,
.access = nfs4_proc_access,
.readlink = nfs4_proc_readlink,
.create = nfs4_proc_create,
.remove = nfs4_proc_remove,
.unlink_setup = nfs4_proc_unlink_setup,
.unlink_done = nfs4_proc_unlink_done,
.rename = nfs4_proc_rename,
.link = nfs4_proc_link,
.symlink = nfs4_proc_symlink,
.mkdir = nfs4_proc_mkdir,
.rmdir = nfs4_proc_remove,
.readdir = nfs4_proc_readdir,
.mknod = nfs4_proc_mknod,
.statfs = nfs4_proc_statfs,
.fsinfo = nfs4_proc_fsinfo,
.pathconf = nfs4_proc_pathconf,
.set_capabilities = nfs4_server_capabilities,
.decode_dirent = nfs4_decode_dirent,
.read_setup = nfs4_proc_read_setup,
.read_done = nfs4_read_done,
.write_setup = nfs4_proc_write_setup,
.write_done = nfs4_write_done,
.commit_setup = nfs4_proc_commit_setup,
.commit_done = nfs4_commit_done,
.file_open = nfs_open,
.file_release = nfs_release,
.lock = nfs4_proc_lock,
.clear_acl_cache = nfs4_zap_acl_attr,
};
/*
* Local variables:
* c-basic-offset: 8
* End:
*/