1
0
mirror of https://git.FreeBSD.org/src.git synced 2024-12-26 11:47:31 +00:00
freebsd/sys/nfsclient/nfs_vfsops.c

1172 lines
33 KiB
C
Raw Normal View History

/*-
* Copyright (c) 1989, 1993, 1995
1994-05-24 10:09:53 +00:00
* The Regents of the University of California. All rights reserved.
*
* This code is derived from software contributed to Berkeley by
* Rick Macklem at The University of Guelph.
*
* 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.
* 4. 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 BY THE REGENTS AND CONTRIBUTORS ``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.
*
* @(#)nfs_vfsops.c 8.12 (Berkeley) 5/20/95
1994-05-24 10:09:53 +00:00
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include "opt_bootp.h"
#include "opt_nfsroot.h"
1994-05-24 10:09:53 +00:00
#include <sys/param.h>
#include <sys/systm.h>
1994-05-24 10:09:53 +00:00
#include <sys/kernel.h>
#include <sys/bio.h>
#include <sys/buf.h>
#include <sys/lock.h>
#include <sys/malloc.h>
1994-05-24 10:09:53 +00:00
#include <sys/mbuf.h>
#include <sys/module.h>
#include <sys/mount.h>
#include <sys/proc.h>
1994-05-24 10:09:53 +00:00
#include <sys/socket.h>
#include <sys/socketvar.h>
#include <sys/sockio.h>
#include <sys/sysctl.h>
#include <sys/vnode.h>
#include <sys/signalvar.h>
1994-05-24 10:09:53 +00:00
#include <vm/vm.h>
#include <vm/vm_extern.h>
#include <vm/uma.h>
1994-05-24 10:09:53 +00:00
#include <net/if.h>
#include <net/route.h>
#include <netinet/in.h>
#include <rpc/rpcclnt.h>
1994-05-24 10:09:53 +00:00
#include <nfs/rpcv2.h>
#include <nfs/nfsproto.h>
#include <nfsclient/nfs.h>
#include <nfsclient/nfsnode.h>
#include <nfsclient/nfsmount.h>
1994-05-24 10:09:53 +00:00
#include <nfs/xdr_subs.h>
#include <nfsclient/nfsm_subs.h>
#include <nfsclient/nfsdiskless.h>
MALLOC_DEFINE(M_NFSREQ, "nfsclient_req", "NFS request header");
MALLOC_DEFINE(M_NFSBIGFH, "nfsclient_bigfh", "NFS version 3 file handle");
MALLOC_DEFINE(M_NFSDIROFF, "nfsclient_diroff", "NFS directory offset data");
MALLOC_DEFINE(M_NFSHASH, "nfsclient_hash", "NFS hash tables");
MALLOC_DEFINE(M_NFSDIRECTIO, "nfsclient_directio", "NFS Direct IO async write state");
uma_zone_t nfsmount_zone;
struct nfsstats nfsstats;
SYSCTL_NODE(_vfs, OID_AUTO, nfs, CTLFLAG_RW, 0, "NFS filesystem");
SYSCTL_STRUCT(_vfs_nfs, NFS_NFSSTATS, nfsstats, CTLFLAG_RW,
&nfsstats, nfsstats, "S,nfsstats");
static int nfs_ip_paranoia = 1;
SYSCTL_INT(_vfs_nfs, OID_AUTO, nfs_ip_paranoia, CTLFLAG_RW,
&nfs_ip_paranoia, 0, "");
#ifdef NFS_DEBUG
int nfs_debug;
SYSCTL_INT(_vfs_nfs, OID_AUTO, debug, CTLFLAG_RW, &nfs_debug, 0, "");
#endif
static int nfs_tprintf_initial_delay = NFS_TPRINTF_INITIAL_DELAY;
SYSCTL_INT(_vfs_nfs, NFS_TPRINTF_INITIAL_DELAY,
downdelayinitial, CTLFLAG_RW, &nfs_tprintf_initial_delay, 0, "");
/* how long between console messages "nfs server foo not responding" */
static int nfs_tprintf_delay = NFS_TPRINTF_DELAY;
SYSCTL_INT(_vfs_nfs, NFS_TPRINTF_DELAY,
downdelayinterval, CTLFLAG_RW, &nfs_tprintf_delay, 0, "");
static void nfs_decode_args(struct mount *mp, struct nfsmount *nmp,
struct nfs_args *argp);
static int mountnfs(struct nfs_args *, struct mount *,
2004-12-07 14:26:39 +00:00
struct sockaddr *, char *, struct vnode **,
struct ucred *cred);
The remaining part of nmount/omount/rootfs mount changes. I cannot sensibly split the conversion of the remaining three filesystems out from the root mounting changes, so in one go: cd9660: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() nfs(client): Convert to nmount (the simple way, mount_nfs(8) is still necessary). Add omount compat shims. Drop COMPAT_PRELITE2 mount arg compatibility. ffs: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() Remove vfs_omount() method, all filesystems are now converted. Remove MNTK_WANTRDWR, handling RO/RW conversions is a filesystem task, and they all do it now. Change rootmounting to use DEVFS trampoline: vfs_mount.c: Mount devfs on /. Devfs needs no 'from' so this is clean. symlink /dev to /. This makes it possible to lookup /dev/foo. Mount "real" root filesystem on /. Surgically move the devfs mountpoint from under the real root filesystem onto /dev in the real root filesystem. Remove now unnecessary getdiskbyname(). kern_init.c: Don't do devfs mounting and rootvnode assignment here, it was already handled by vfs_mount.c. Remove now unused bdevvp(), addaliasu() and addalias(). Put the few necessary lines in devfs where they belong. This eliminates the second-last source of bogo vnodes, leaving only the lemming-syncer. Remove rootdev variable, it doesn't give meaning in a global context and was not trustworth anyway. Correct information is provided by statfs(/).
2004-12-07 08:15:41 +00:00
static vfs_mount_t nfs_mount;
static vfs_cmount_t nfs_cmount;
static vfs_unmount_t nfs_unmount;
static vfs_root_t nfs_root;
static vfs_statfs_t nfs_statfs;
static vfs_sync_t nfs_sync;
static vfs_sysctl_t nfs_sysctl;
1994-10-23 23:26:18 +00:00
1994-05-24 10:09:53 +00:00
/*
* nfs vfs operations.
*/
1995-12-17 21:14:36 +00:00
static struct vfsops nfs_vfsops = {
.vfs_init = nfs_init,
The remaining part of nmount/omount/rootfs mount changes. I cannot sensibly split the conversion of the remaining three filesystems out from the root mounting changes, so in one go: cd9660: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() nfs(client): Convert to nmount (the simple way, mount_nfs(8) is still necessary). Add omount compat shims. Drop COMPAT_PRELITE2 mount arg compatibility. ffs: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() Remove vfs_omount() method, all filesystems are now converted. Remove MNTK_WANTRDWR, handling RO/RW conversions is a filesystem task, and they all do it now. Change rootmounting to use DEVFS trampoline: vfs_mount.c: Mount devfs on /. Devfs needs no 'from' so this is clean. symlink /dev to /. This makes it possible to lookup /dev/foo. Mount "real" root filesystem on /. Surgically move the devfs mountpoint from under the real root filesystem onto /dev in the real root filesystem. Remove now unnecessary getdiskbyname(). kern_init.c: Don't do devfs mounting and rootvnode assignment here, it was already handled by vfs_mount.c. Remove now unused bdevvp(), addaliasu() and addalias(). Put the few necessary lines in devfs where they belong. This eliminates the second-last source of bogo vnodes, leaving only the lemming-syncer. Remove rootdev variable, it doesn't give meaning in a global context and was not trustworth anyway. Correct information is provided by statfs(/).
2004-12-07 08:15:41 +00:00
.vfs_mount = nfs_mount,
.vfs_cmount = nfs_cmount,
.vfs_root = nfs_root,
.vfs_statfs = nfs_statfs,
.vfs_sync = nfs_sync,
.vfs_uninit = nfs_uninit,
.vfs_unmount = nfs_unmount,
.vfs_sysctl = nfs_sysctl,
1994-05-24 10:09:53 +00:00
};
VFS_SET(nfs_vfsops, nfs, VFCF_NETWORK);
1994-05-24 10:09:53 +00:00
/* So that loader and kldload(2) can find us, wherever we are.. */
MODULE_VERSION(nfs, 1);
static struct nfs_rpcops nfs_rpcops = {
nfs_readrpc,
nfs_writerpc,
nfs_writebp,
nfs_readlinkrpc,
nfs_invaldir,
nfs_commit,
};
1994-05-24 10:09:53 +00:00
/*
* This structure must be filled in by a primary bootstrap or bootstrap
* server for a diskless/dataless machine. It is initialized below just
* to ensure that it is allocated to initialized data (.data not .bss).
*/
struct nfs_diskless nfs_diskless = { { { 0 } } };
struct nfsv3_diskless nfsv3_diskless = { { { 0 } } };
int nfs_diskless_valid = 0;
1994-05-24 10:09:53 +00:00
SYSCTL_INT(_vfs_nfs, OID_AUTO, diskless_valid, CTLFLAG_RD,
&nfs_diskless_valid, 0, "");
SYSCTL_STRING(_vfs_nfs, OID_AUTO, diskless_rootpath, CTLFLAG_RD,
nfsv3_diskless.root_hostnam, 0, "");
SYSCTL_OPAQUE(_vfs_nfs, OID_AUTO, diskless_rootaddr, CTLFLAG_RD,
&nfsv3_diskless.root_saddr, sizeof nfsv3_diskless.root_saddr,
"%Ssockaddr_in", "");
void nfsargs_ntoh(struct nfs_args *);
static int nfs_mountdiskless(char *,
struct sockaddr_in *, struct nfs_args *,
struct thread *, struct vnode **, struct mount *);
static void nfs_convert_diskless(void);
static void nfs_convert_oargs(struct nfs_args *args,
struct onfs_args *oargs);
1994-05-24 10:09:53 +00:00
int
nfs_iosize(struct nfsmount *nmp)
{
int iosize;
/*
* Calculate the size used for io buffers. Use the larger
* of the two sizes to minimise nfs requests but make sure
* that it is at least one VM page to avoid wasting buffer
* space.
*/
iosize = imax(nmp->nm_rsize, nmp->nm_wsize);
iosize = imax(iosize, PAGE_SIZE);
return (iosize);
}
static void
nfs_convert_oargs(struct nfs_args *args, struct onfs_args *oargs)
{
args->version = NFS_ARGSVERSION;
args->addr = oargs->addr;
args->addrlen = oargs->addrlen;
args->sotype = oargs->sotype;
args->proto = oargs->proto;
args->fh = oargs->fh;
args->fhsize = oargs->fhsize;
args->flags = oargs->flags;
args->wsize = oargs->wsize;
args->rsize = oargs->rsize;
args->readdirsize = oargs->readdirsize;
args->timeo = oargs->timeo;
args->retrans = oargs->retrans;
args->maxgrouplist = oargs->maxgrouplist;
args->readahead = oargs->readahead;
args->deadthresh = oargs->deadthresh;
args->hostname = oargs->hostname;
}
1998-02-09 06:11:36 +00:00
static void
nfs_convert_diskless(void)
{
bcopy(&nfs_diskless.myif, &nfsv3_diskless.myif,
sizeof(struct ifaliasreq));
bcopy(&nfs_diskless.mygateway, &nfsv3_diskless.mygateway,
sizeof(struct sockaddr_in));
nfs_convert_oargs(&nfsv3_diskless.root_args,&nfs_diskless.root_args);
if (nfsv3_diskless.root_args.flags & NFSMNT_NFSV3) {
nfsv3_diskless.root_fhsize = NFSX_V3FH;
bcopy(nfs_diskless.root_fh, nfsv3_diskless.root_fh, NFSX_V3FH);
} else {
nfsv3_diskless.root_fhsize = NFSX_V2FH;
bcopy(nfs_diskless.root_fh, nfsv3_diskless.root_fh, NFSX_V2FH);
}
bcopy(&nfs_diskless.root_saddr,&nfsv3_diskless.root_saddr,
sizeof(struct sockaddr_in));
bcopy(nfs_diskless.root_hostnam, nfsv3_diskless.root_hostnam, MNAMELEN);
nfsv3_diskless.root_time = nfs_diskless.root_time;
bcopy(nfs_diskless.my_hostnam, nfsv3_diskless.my_hostnam,
MAXHOSTNAMELEN);
nfs_diskless_valid = 3;
}
1994-05-24 10:09:53 +00:00
/*
* nfs statfs call
*/
static int
nfs_statfs(struct mount *mp, struct statfs *sbp, struct thread *td)
1994-05-24 10:09:53 +00:00
{
struct vnode *vp;
struct nfs_statfs *sfp;
caddr_t bpos, dpos;
struct nfsmount *nmp = VFSTONFS(mp);
int error = 0, v3 = (nmp->nm_flag & NFSMNT_NFSV3), retattr;
struct mbuf *mreq, *mrep, *md, *mb;
1994-05-24 10:09:53 +00:00
struct nfsnode *np;
u_quad_t tquad;
1994-05-24 10:09:53 +00:00
#ifndef nolint
sfp = NULL;
#endif
error = vfs_busy(mp, LK_NOWAIT, NULL, td);
if (error)
1994-05-24 10:09:53 +00:00
return (error);
error = nfs_nget(mp, (nfsfh_t *)nmp->nm_fh, nmp->nm_fhsize, &np, LK_EXCLUSIVE);
if (error) {
vfs_unbusy(mp, td);
return (error);
}
1994-05-24 10:09:53 +00:00
vp = NFSTOV(np);
mtx_lock(&nmp->nm_mtx);
if (v3 && (nmp->nm_state & NFSSTA_GOTFSINFO) == 0) {
mtx_unlock(&nmp->nm_mtx);
(void)nfs_fsinfo(nmp, vp, td->td_ucred, td);
} else
mtx_unlock(&nmp->nm_mtx);
nfsstats.rpccnt[NFSPROC_FSSTAT]++;
mreq = nfsm_reqhead(vp, NFSPROC_FSSTAT, NFSX_FH(v3));
mb = mreq;
bpos = mtod(mb, caddr_t);
nfsm_fhtom(vp, v3);
nfsm_request(vp, NFSPROC_FSSTAT, td, td->td_ucred);
if (v3)
nfsm_postop_attr(vp, retattr);
if (error) {
if (mrep != NULL)
m_freem(mrep);
Fix a condition where nfs_statfs() can precipitate a panic. There is code that says this: nfsm_request(vp, NFSPROC_FSSTAT, p, cred); if (v3) nfsm_postop_attr(vp, retattr); if (!error) nfsm_dissect(sfp, struct nfs_statfs *, NFSX_STATFS(v3)); The problem here is that if error != 0, nfsm_dissect() will not be called, which leaves sfp == NULL. But nfs_statfs() does not bail out at this point: it continues processing until it tries to dereference sfp, which causes a panic. I was able to generate this crash under the following conditions: 1) Set up a machine as an NFS server and NFS client, with amd running (using NIS maps). /usr/local is exported, though any exported fs can can be used to trigger the bug. 2) Log in as normal user, with home directory mounted from a SunOS 4.1.3 NFS server via amd (along with a few other NFS filesystems from same machine). 3) Su to root and type the following: # mount localhost:/usr/local /mnt # df To fix the panic, I changed the code to read: if (!error) { nfsm_dissect(sfp, struct nfs_statfs *, NFSX_STATFS(v3)); } else goto nfsmout; This is a bit kludgy in that nfsmout is a label defined by the nfsm_subs.h macros, but these macros are themselves more than a little kludgy. This stops the machine from crashing, but does not fix the overall bug: 'error' somehow becomes 5 (EIO) when a statfs() is performed on the locally mounted NFS filesystem. This seems to only happen the first time the filesystem is accesed: on subsequent accesses, it seems to work fine again. Now, I know there's no practical use in mounting a local filesystem via NFS, but doing it shouldn't cause the system to melt down.
1997-06-27 19:10:46 +00:00
goto nfsmout;
}
sfp = nfsm_dissect(struct nfs_statfs *, NFSX_STATFS(v3));
mtx_lock(&nmp->nm_mtx);
sbp->f_iosize = nfs_iosize(nmp);
mtx_unlock(&nmp->nm_mtx);
if (v3) {
sbp->f_bsize = NFS_FABLKSIZE;
tquad = fxdr_hyper(&sfp->sf_tbytes);
sbp->f_blocks = tquad / NFS_FABLKSIZE;
tquad = fxdr_hyper(&sfp->sf_fbytes);
sbp->f_bfree = tquad / NFS_FABLKSIZE;
tquad = fxdr_hyper(&sfp->sf_abytes);
sbp->f_bavail = tquad / NFS_FABLKSIZE;
sbp->f_files = (fxdr_unsigned(int32_t,
sfp->sf_tfiles.nfsuquad[1]) & 0x7fffffff);
sbp->f_ffree = (fxdr_unsigned(int32_t,
sfp->sf_ffiles.nfsuquad[1]) & 0x7fffffff);
1994-05-24 10:09:53 +00:00
} else {
sbp->f_bsize = fxdr_unsigned(int32_t, sfp->sf_bsize);
sbp->f_blocks = fxdr_unsigned(int32_t, sfp->sf_blocks);
sbp->f_bfree = fxdr_unsigned(int32_t, sfp->sf_bfree);
sbp->f_bavail = fxdr_unsigned(int32_t, sfp->sf_bavail);
1994-05-24 10:09:53 +00:00
sbp->f_files = 0;
sbp->f_ffree = 0;
}
m_freem(mrep);
nfsmout:
vput(vp);
vfs_unbusy(mp, td);
1994-05-24 10:09:53 +00:00
return (error);
}
/*
* nfs version 3 fsinfo rpc call
*/
int
nfs_fsinfo(struct nfsmount *nmp, struct vnode *vp, struct ucred *cred,
struct thread *td)
{
struct nfsv3_fsinfo *fsp;
u_int32_t pref, max;
caddr_t bpos, dpos;
int error = 0, retattr;
struct mbuf *mreq, *mrep, *md, *mb;
u_int64_t maxfsize;
nfsstats.rpccnt[NFSPROC_FSINFO]++;
mreq = nfsm_reqhead(vp, NFSPROC_FSINFO, NFSX_FH(1));
mb = mreq;
bpos = mtod(mb, caddr_t);
nfsm_fhtom(vp, 1);
nfsm_request(vp, NFSPROC_FSINFO, td, cred);
nfsm_postop_attr(vp, retattr);
if (!error) {
fsp = nfsm_dissect(struct nfsv3_fsinfo *, NFSX_V3FSINFO);
pref = fxdr_unsigned(u_int32_t, fsp->fs_wtpref);
mtx_lock(&nmp->nm_mtx);
if (pref < nmp->nm_wsize && pref >= NFS_FABLKSIZE)
nmp->nm_wsize = (pref + NFS_FABLKSIZE - 1) &
~(NFS_FABLKSIZE - 1);
max = fxdr_unsigned(u_int32_t, fsp->fs_wtmax);
if (max < nmp->nm_wsize && max > 0) {
nmp->nm_wsize = max & ~(NFS_FABLKSIZE - 1);
if (nmp->nm_wsize == 0)
nmp->nm_wsize = max;
}
pref = fxdr_unsigned(u_int32_t, fsp->fs_rtpref);
if (pref < nmp->nm_rsize && pref >= NFS_FABLKSIZE)
nmp->nm_rsize = (pref + NFS_FABLKSIZE - 1) &
~(NFS_FABLKSIZE - 1);
max = fxdr_unsigned(u_int32_t, fsp->fs_rtmax);
if (max < nmp->nm_rsize && max > 0) {
nmp->nm_rsize = max & ~(NFS_FABLKSIZE - 1);
if (nmp->nm_rsize == 0)
nmp->nm_rsize = max;
}
pref = fxdr_unsigned(u_int32_t, fsp->fs_dtpref);
if (pref < nmp->nm_readdirsize && pref >= NFS_DIRBLKSIZ)
nmp->nm_readdirsize = (pref + NFS_DIRBLKSIZ - 1) &
~(NFS_DIRBLKSIZ - 1);
if (max < nmp->nm_readdirsize && max > 0) {
nmp->nm_readdirsize = max & ~(NFS_DIRBLKSIZ - 1);
if (nmp->nm_readdirsize == 0)
nmp->nm_readdirsize = max;
}
maxfsize = fxdr_hyper(&fsp->fs_maxfilesize);
if (maxfsize > 0 && maxfsize < nmp->nm_maxfilesize)
nmp->nm_maxfilesize = maxfsize;
nmp->nm_mountp->mnt_stat.f_iosize = nfs_iosize(nmp);
nmp->nm_state |= NFSSTA_GOTFSINFO;
mtx_unlock(&nmp->nm_mtx);
}
m_freem(mrep);
nfsmout:
return (error);
}
1994-05-24 10:09:53 +00:00
/*
* Mount a remote root fs via. nfs. This depends on the info in the
* nfs_diskless structure that has been filled in properly by some primary
* bootstrap.
* It goes something like this:
* - do enough of "ifconfig" by calling ifioctl() so that the system
* can talk to the server
* - If nfs_diskless.mygateway is filled in, use that address as
* a default gateway.
* - build the rootfs mount point and call mountnfs() to do the rest.
*
* It is assumed to be safe to read, modify, and write the nfsv3_diskless
* structure, as well as other global NFS client variables here, as
* nfs_mountroot() will be called once in the boot before any other NFS
* client activity occurs.
1994-05-24 10:09:53 +00:00
*/
int
nfs_mountroot(struct mount *mp, struct thread *td)
1994-05-24 10:09:53 +00:00
{
struct nfsv3_diskless *nd = &nfsv3_diskless;
1994-05-24 10:09:53 +00:00
struct socket *so;
struct vnode *vp;
struct ifreq ir;
1994-05-24 10:09:53 +00:00
int error, i;
u_long l;
char buf[128];
char *cp;
1994-05-24 10:09:53 +00:00
#if defined(BOOTP_NFSROOT) && defined(BOOTP)
bootpc_init(); /* use bootp to get nfs_diskless filled in */
#elif defined(NFS_ROOT)
nfs_setup_diskless();
#endif
if (nfs_diskless_valid == 0)
return (-1);
if (nfs_diskless_valid == 1)
nfs_convert_diskless();
1995-05-30 08:16:23 +00:00
/*
* XXX splnet, so networks will receive...
*/
splnet();
1994-05-24 10:09:53 +00:00
/*
* Do enough of ifconfig(8) so that the critical net interface can
* talk to the server.
*/
error = socreate(nd->myif.ifra_addr.sa_family, &so, nd->root_args.sotype, 0,
td->td_ucred, td);
if (error)
panic("nfs_mountroot: socreate(%04x): %d",
nd->myif.ifra_addr.sa_family, error);
#if 0 /* XXX Bad idea */
/*
* We might not have been told the right interface, so we pass
* over the first ten interfaces of the same kind, until we get
* one of them configured.
*/
for (i = strlen(nd->myif.ifra_name) - 1;
1995-05-30 08:16:23 +00:00
nd->myif.ifra_name[i] >= '0' &&
nd->myif.ifra_name[i] <= '9';
nd->myif.ifra_name[i] ++) {
error = ifioctl(so, SIOCAIFADDR, (caddr_t)&nd->myif, td);
1995-05-30 08:16:23 +00:00
if(!error)
break;
}
#endif
error = ifioctl(so, SIOCAIFADDR, (caddr_t)&nd->myif, td);
if (error)
1994-05-24 10:09:53 +00:00
panic("nfs_mountroot: SIOCAIFADDR: %d", error);
if ((cp = getenv("boot.netif.mtu")) != NULL) {
ir.ifr_mtu = strtol(cp, NULL, 10);
bcopy(nd->myif.ifra_name, ir.ifr_name, IFNAMSIZ);
freeenv(cp);
error = ifioctl(so, SIOCSIFMTU, (caddr_t)&ir, td);
if (error)
printf("nfs_mountroot: SIOCSIFMTU: %d", error);
}
1994-05-24 10:09:53 +00:00
soclose(so);
/*
* If the gateway field is filled in, set it as the default route.
* Note that pxeboot will set a default route of 0 if the route
* is not set by the DHCP server. Check also for a value of 0
* to avoid panicking inappropriately in that situation.
1994-05-24 10:09:53 +00:00
*/
if (nd->mygateway.sin_len != 0 &&
nd->mygateway.sin_addr.s_addr != 0) {
1994-05-24 10:09:53 +00:00
struct sockaddr_in mask, sin;
bzero((caddr_t)&mask, sizeof(mask));
sin = mask;
sin.sin_family = AF_INET;
sin.sin_len = sizeof(sin);
Add code to allow the system to handle multiple routing tables. This particular implementation is designed to be fully backwards compatible and to be MFC-able to 7.x (and 6.x) Currently the only protocol that can make use of the multiple tables is IPv4 Similar functionality exists in OpenBSD and Linux. From my notes: ----- One thing where FreeBSD has been falling behind, and which by chance I have some time to work on is "policy based routing", which allows different packet streams to be routed by more than just the destination address. Constraints: ------------ I want to make some form of this available in the 6.x tree (and by extension 7.x) , but FreeBSD in general needs it so I might as well do it in -current and back port the portions I need. One of the ways that this can be done is to have the ability to instantiate multiple kernel routing tables (which I will now refer to as "Forwarding Information Bases" or "FIBs" for political correctness reasons). Which FIB a particular packet uses to make the next hop decision can be decided by a number of mechanisms. The policies these mechanisms implement are the "Policies" referred to in "Policy based routing". One of the constraints I have if I try to back port this work to 6.x is that it must be implemented as a EXTENSION to the existing ABIs in 6.x so that third party applications do not need to be recompiled in timespan of the branch. This first version will not have some of the bells and whistles that will come with later versions. It will, for example, be limited to 16 tables in the first commit. Implementation method, Compatible version. (part 1) ------------------------------- For this reason I have implemented a "sufficient subset" of a multiple routing table solution in Perforce, and back-ported it to 6.x. (also in Perforce though not always caught up with what I have done in -current/P4). The subset allows a number of FIBs to be defined at compile time (8 is sufficient for my purposes in 6.x) and implements the changes needed to allow IPV4 to use them. I have not done the changes for ipv6 simply because I do not need it, and I do not have enough knowledge of ipv6 (e.g. neighbor discovery) needed to do it. Other protocol families are left untouched and should there be users with proprietary protocol families, they should continue to work and be oblivious to the existence of the extra FIBs. To understand how this is done, one must know that the current FIB code starts everything off with a single dimensional array of pointers to FIB head structures (One per protocol family), each of which in turn points to the trie of routes available to that family. The basic change in the ABI compatible version of the change is to extent that array to be a 2 dimensional array, so that instead of protocol family X looking at rt_tables[X] for the table it needs, it looks at rt_tables[Y][X] when for all protocol families except ipv4 Y is always 0. Code that is unaware of the change always just sees the first row of the table, which of course looks just like the one dimensional array that existed before. The entry points rtrequest(), rtalloc(), rtalloc1(), rtalloc_ign() are all maintained, but refer only to the first row of the array, so that existing callers in proprietary protocols can continue to do the "right thing". Some new entry points are added, for the exclusive use of ipv4 code called in_rtrequest(), in_rtalloc(), in_rtalloc1() and in_rtalloc_ign(), which have an extra argument which refers the code to the correct row. In addition, there are some new entry points (currently called rtalloc_fib() and friends) that check the Address family being looked up and call either rtalloc() (and friends) if the protocol is not IPv4 forcing the action to row 0 or to the appropriate row if it IS IPv4 (and that info is available). These are for calling from code that is not specific to any particular protocol. The way these are implemented would change in the non ABI preserving code to be added later. One feature of the first version of the code is that for ipv4, the interface routes show up automatically on all the FIBs, so that no matter what FIB you select you always have the basic direct attached hosts available to you. (rtinit() does this automatically). You CAN delete an interface route from one FIB should you want to but by default it's there. ARP information is also available in each FIB. It's assumed that the same machine would have the same MAC address, regardless of which FIB you are using to get to it. This brings us as to how the correct FIB is selected for an outgoing IPV4 packet. Firstly, all packets have a FIB associated with them. if nothing has been done to change it, it will be FIB 0. The FIB is changed in the following ways. Packets fall into one of a number of classes. 1/ locally generated packets, coming from a socket/PCB. Such packets select a FIB from a number associated with the socket/PCB. This in turn is inherited from the process, but can be changed by a socket option. The process in turn inherits it on fork. I have written a utility call setfib that acts a bit like nice.. setfib -3 ping target.example.com # will use fib 3 for ping. It is an obvious extension to make it a property of a jail but I have not done so. It can be achieved by combining the setfib and jail commands. 2/ packets received on an interface for forwarding. By default these packets would use table 0, (or possibly a number settable in a sysctl(not yet)). but prior to routing the firewall can inspect them (see below). (possibly in the future you may be able to associate a FIB with packets received on an interface.. An ifconfig arg, but not yet.) 3/ packets inspected by a packet classifier, which can arbitrarily associate a fib with it on a packet by packet basis. A fib assigned to a packet by a packet classifier (such as ipfw) would over-ride a fib associated by a more default source. (such as cases 1 or 2). 4/ a tcp listen socket associated with a fib will generate accept sockets that are associated with that same fib. 5/ Packets generated in response to some other packet (e.g. reset or icmp packets). These should use the FIB associated with the packet being reponded to. 6/ Packets generated during encapsulation. gif, tun and other tunnel interfaces will encapsulate using the FIB that was in effect withthe proces that set up the tunnel. thus setfib 1 ifconfig gif0 [tunnel instructions] will set the fib for the tunnel to use to be fib 1. Routing messages would be associated with their process, and thus select one FIB or another. messages from the kernel would be associated with the fib they refer to and would only be received by a routing socket associated with that fib. (not yet implemented) In addition Netstat has been edited to be able to cope with the fact that the array is now 2 dimensional. (It looks in system memory using libkvm (!)). Old versions of netstat see only the first FIB. In addition two sysctls are added to give: a) the number of FIBs compiled in (active) b) the default FIB of the calling process. Early testing experience: ------------------------- Basically our (IronPort's) appliance does this functionality already using ipfw fwd but that method has some drawbacks. For example, It can't fully simulate a routing table because it can't influence the socket's choice of local address when a connect() is done. Testing during the generating of these changes has been remarkably smooth so far. Multiple tables have co-existed with no notable side effects, and packets have been routes accordingly. ipfw has grown 2 new keywords: setfib N ip from anay to any count ip from any to any fib N In pf there seems to be a requirement to be able to give symbolic names to the fibs but I do not have that capacity. I am not sure if it is required. SCTP has interestingly enough built in support for this, called VRFs in Cisco parlance. it will be interesting to see how that handles it when it suddenly actually does something. Where to next: -------------------- After committing the ABI compatible version and MFCing it, I'd like to proceed in a forward direction in -current. this will result in some roto-tilling in the routing code. Firstly: the current code's idea of having a separate tree per protocol family, all of the same format, and pointed to by the 1 dimensional array is a bit silly. Especially when one considers that there is code that makes assumptions about every protocol having the same internal structures there. Some protocols don't WANT that sort of structure. (for example the whole idea of a netmask is foreign to appletalk). This needs to be made opaque to the external code. My suggested first change is to add routing method pointers to the 'domain' structure, along with information pointing the data. instead of having an array of pointers to uniform structures, there would be an array pointing to the 'domain' structures for each protocol address domain (protocol family), and the methods this reached would be called. The methods would have an argument that gives FIB number, but the protocol would be free to ignore it. When the ABI can be changed it raises the possibilty of the addition of a fib entry into the "struct route". Currently, the structure contains the sockaddr of the desination, and the resulting fib entry. To make this work fully, one could add a fib number so that given an address and a fib, one can find the third element, the fib entry. Interaction with the ARP layer/ LL layer would need to be revisited as well. Qing Li has been working on this already. This work was sponsored by Ironport Systems/Cisco Reviewed by: several including rwatson, bz and mlair (parts each) Obtained from: Ironport systems/Cisco
2008-05-09 23:03:00 +00:00
/* XXX MRT use table 0 for this sort of thing */
error = rtrequest(RTM_ADD, (struct sockaddr *)&sin,
1994-05-24 10:09:53 +00:00
(struct sockaddr *)&nd->mygateway,
(struct sockaddr *)&mask,
RTF_UP | RTF_GATEWAY, NULL);
if (error)
1994-05-24 10:09:53 +00:00
panic("nfs_mountroot: RTM_ADD: %d", error);
}
/*
* Create the rootfs mount point.
*/
nd->root_args.fh = nd->root_fh;
nd->root_args.fhsize = nd->root_fhsize;
l = ntohl(nd->root_saddr.sin_addr.s_addr);
snprintf(buf, sizeof(buf), "%ld.%ld.%ld.%ld:%s",
(l >> 24) & 0xff, (l >> 16) & 0xff,
(l >> 8) & 0xff, (l >> 0) & 0xff, nd->root_hostnam);
printf("NFS ROOT: %s\n", buf);
nd->root_args.hostname = buf;
if ((error = nfs_mountdiskless(buf,
&nd->root_saddr, &nd->root_args, td, &vp, mp)) != 0) {
return (error);
}
1994-05-24 10:09:53 +00:00
/*
* This is not really an nfs issue, but it is much easier to
* set hostname here and then let the "/etc/rc.xxx" files
* mount the right /var based upon its preset value.
*/
bcopy(nd->my_hostnam, hostname, MAXHOSTNAMELEN);
hostname[MAXHOSTNAMELEN - 1] = '\0';
for (i = 0; i < MAXHOSTNAMELEN; i++)
if (hostname[i] == '\0')
break;
inittodr(ntohl(nd->root_time));
return (0);
}
/*
* Internal version of mount system call for diskless setup.
*/
static int
nfs_mountdiskless(char *path,
struct sockaddr_in *sin, struct nfs_args *args, struct thread *td,
struct vnode **vpp, struct mount *mp)
1994-05-24 10:09:53 +00:00
{
struct sockaddr *nam;
int error;
nam = sodupsockaddr((struct sockaddr *)sin, M_WAITOK);
2004-12-07 14:26:39 +00:00
if ((error = mountnfs(args, mp, nam, path, vpp,
td->td_ucred)) != 0) {
2004-12-07 14:26:39 +00:00
printf("nfs_mountroot: mount %s on /: %d\n", path, error);
return (error);
}
return (0);
1994-05-24 10:09:53 +00:00
}
static void
The remaining part of nmount/omount/rootfs mount changes. I cannot sensibly split the conversion of the remaining three filesystems out from the root mounting changes, so in one go: cd9660: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() nfs(client): Convert to nmount (the simple way, mount_nfs(8) is still necessary). Add omount compat shims. Drop COMPAT_PRELITE2 mount arg compatibility. ffs: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() Remove vfs_omount() method, all filesystems are now converted. Remove MNTK_WANTRDWR, handling RO/RW conversions is a filesystem task, and they all do it now. Change rootmounting to use DEVFS trampoline: vfs_mount.c: Mount devfs on /. Devfs needs no 'from' so this is clean. symlink /dev to /. This makes it possible to lookup /dev/foo. Mount "real" root filesystem on /. Surgically move the devfs mountpoint from under the real root filesystem onto /dev in the real root filesystem. Remove now unnecessary getdiskbyname(). kern_init.c: Don't do devfs mounting and rootvnode assignment here, it was already handled by vfs_mount.c. Remove now unused bdevvp(), addaliasu() and addalias(). Put the few necessary lines in devfs where they belong. This eliminates the second-last source of bogo vnodes, leaving only the lemming-syncer. Remove rootdev variable, it doesn't give meaning in a global context and was not trustworth anyway. Correct information is provided by statfs(/).
2004-12-07 08:15:41 +00:00
nfs_decode_args(struct mount *mp, struct nfsmount *nmp, struct nfs_args *argp)
{
int s;
int adjsock;
int maxio;
char *p;
s = splnet();
/*
* Set read-only flag if requested; otherwise, clear it if this is
* an update. If this is not an update, then either the read-only
* flag is already clear, or this is a root mount and it was set
* intentionally at some previous point.
*/
if (vfs_getopt(mp->mnt_optnew, "ro", NULL, NULL) == 0) {
MNT_ILOCK(mp);
The remaining part of nmount/omount/rootfs mount changes. I cannot sensibly split the conversion of the remaining three filesystems out from the root mounting changes, so in one go: cd9660: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() nfs(client): Convert to nmount (the simple way, mount_nfs(8) is still necessary). Add omount compat shims. Drop COMPAT_PRELITE2 mount arg compatibility. ffs: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() Remove vfs_omount() method, all filesystems are now converted. Remove MNTK_WANTRDWR, handling RO/RW conversions is a filesystem task, and they all do it now. Change rootmounting to use DEVFS trampoline: vfs_mount.c: Mount devfs on /. Devfs needs no 'from' so this is clean. symlink /dev to /. This makes it possible to lookup /dev/foo. Mount "real" root filesystem on /. Surgically move the devfs mountpoint from under the real root filesystem onto /dev in the real root filesystem. Remove now unnecessary getdiskbyname(). kern_init.c: Don't do devfs mounting and rootvnode assignment here, it was already handled by vfs_mount.c. Remove now unused bdevvp(), addaliasu() and addalias(). Put the few necessary lines in devfs where they belong. This eliminates the second-last source of bogo vnodes, leaving only the lemming-syncer. Remove rootdev variable, it doesn't give meaning in a global context and was not trustworth anyway. Correct information is provided by statfs(/).
2004-12-07 08:15:41 +00:00
mp->mnt_flag |= MNT_RDONLY;
MNT_IUNLOCK(mp);
} else if (mp->mnt_flag & MNT_UPDATE) {
MNT_ILOCK(mp);
mp->mnt_flag &= ~MNT_RDONLY;
MNT_IUNLOCK(mp);
}
/*
* Silently clear NFSMNT_NOCONN if it's a TCP mount, it makes
* no sense in that context. Also, set up appropriate retransmit
* and soft timeout behavior.
*/
if (argp->sotype == SOCK_STREAM) {
nmp->nm_flag &= ~NFSMNT_NOCONN;
nmp->nm_flag |= NFSMNT_DUMBTIMR;
nmp->nm_timeo = NFS_MAXTIMEO;
nmp->nm_retry = NFS_RETRANS_TCP;
}
/* Also clear RDIRPLUS if not NFSv3, it crashes some servers */
if ((argp->flags & NFSMNT_NFSV3) == 0)
nmp->nm_flag &= ~NFSMNT_RDIRPLUS;
/* Re-bind if rsrvd port requested and wasn't on one */
adjsock = !(nmp->nm_flag & NFSMNT_RESVPORT)
&& (argp->flags & NFSMNT_RESVPORT);
/* Also re-bind if we're switching to/from a connected UDP socket */
adjsock |= ((nmp->nm_flag & NFSMNT_NOCONN) !=
(argp->flags & NFSMNT_NOCONN));
/* Update flags atomically. Don't change the lock bits. */
nmp->nm_flag = argp->flags | nmp->nm_flag;
splx(s);
if ((argp->flags & NFSMNT_TIMEO) && argp->timeo > 0) {
nmp->nm_timeo = (argp->timeo * NFS_HZ + 5) / 10;
if (nmp->nm_timeo < NFS_MINTIMEO)
nmp->nm_timeo = NFS_MINTIMEO;
else if (nmp->nm_timeo > NFS_MAXTIMEO)
nmp->nm_timeo = NFS_MAXTIMEO;
}
if ((argp->flags & NFSMNT_RETRANS) && argp->retrans > 1) {
nmp->nm_retry = argp->retrans;
if (nmp->nm_retry > NFS_MAXREXMIT)
nmp->nm_retry = NFS_MAXREXMIT;
}
if (argp->flags & NFSMNT_NFSV3) {
if (argp->sotype == SOCK_DGRAM)
maxio = NFS_MAXDGRAMDATA;
else
maxio = NFS_MAXDATA;
} else
maxio = NFS_V2MAXDATA;
if ((argp->flags & NFSMNT_WSIZE) && argp->wsize > 0) {
nmp->nm_wsize = argp->wsize;
/* Round down to multiple of blocksize */
nmp->nm_wsize &= ~(NFS_FABLKSIZE - 1);
if (nmp->nm_wsize <= 0)
nmp->nm_wsize = NFS_FABLKSIZE;
}
if (nmp->nm_wsize > maxio)
nmp->nm_wsize = maxio;
if (nmp->nm_wsize > MAXBSIZE)
nmp->nm_wsize = MAXBSIZE;
if ((argp->flags & NFSMNT_RSIZE) && argp->rsize > 0) {
nmp->nm_rsize = argp->rsize;
/* Round down to multiple of blocksize */
nmp->nm_rsize &= ~(NFS_FABLKSIZE - 1);
if (nmp->nm_rsize <= 0)
nmp->nm_rsize = NFS_FABLKSIZE;
}
if (nmp->nm_rsize > maxio)
nmp->nm_rsize = maxio;
if (nmp->nm_rsize > MAXBSIZE)
nmp->nm_rsize = MAXBSIZE;
if ((argp->flags & NFSMNT_READDIRSIZE) && argp->readdirsize > 0) {
nmp->nm_readdirsize = argp->readdirsize;
}
if (nmp->nm_readdirsize > maxio)
nmp->nm_readdirsize = maxio;
if (nmp->nm_readdirsize > nmp->nm_rsize)
nmp->nm_readdirsize = nmp->nm_rsize;
if ((argp->flags & NFSMNT_ACREGMIN) && argp->acregmin >= 0)
nmp->nm_acregmin = argp->acregmin;
else
nmp->nm_acregmin = NFS_MINATTRTIMO;
if ((argp->flags & NFSMNT_ACREGMAX) && argp->acregmax >= 0)
nmp->nm_acregmax = argp->acregmax;
else
nmp->nm_acregmax = NFS_MAXATTRTIMO;
if ((argp->flags & NFSMNT_ACDIRMIN) && argp->acdirmin >= 0)
nmp->nm_acdirmin = argp->acdirmin;
else
nmp->nm_acdirmin = NFS_MINDIRATTRTIMO;
if ((argp->flags & NFSMNT_ACDIRMAX) && argp->acdirmax >= 0)
nmp->nm_acdirmax = argp->acdirmax;
else
nmp->nm_acdirmax = NFS_MAXDIRATTRTIMO;
if (nmp->nm_acdirmin > nmp->nm_acdirmax)
nmp->nm_acdirmin = nmp->nm_acdirmax;
if (nmp->nm_acregmin > nmp->nm_acregmax)
nmp->nm_acregmin = nmp->nm_acregmax;
if ((argp->flags & NFSMNT_MAXGRPS) && argp->maxgrouplist >= 0) {
if (argp->maxgrouplist <= NFS_MAXGRPS)
nmp->nm_numgrps = argp->maxgrouplist;
else
nmp->nm_numgrps = NFS_MAXGRPS;
}
if ((argp->flags & NFSMNT_READAHEAD) && argp->readahead >= 0) {
if (argp->readahead <= NFS_MAXRAHEAD)
nmp->nm_readahead = argp->readahead;
else
nmp->nm_readahead = NFS_MAXRAHEAD;
}
if ((argp->flags & NFSMNT_WCOMMITSIZE) && argp->wcommitsize >= 0) {
if (argp->wcommitsize < nmp->nm_wsize)
nmp->nm_wcommitsize = nmp->nm_wsize;
else
nmp->nm_wcommitsize = argp->wcommitsize;
}
if ((argp->flags & NFSMNT_DEADTHRESH) && argp->deadthresh >= 0) {
if (argp->deadthresh <= NFS_MAXDEADTHRESH)
nmp->nm_deadthresh = argp->deadthresh;
else
nmp->nm_deadthresh = NFS_MAXDEADTHRESH;
}
adjsock |= ((nmp->nm_sotype != argp->sotype) ||
(nmp->nm_soproto != argp->proto));
nmp->nm_sotype = argp->sotype;
nmp->nm_soproto = argp->proto;
if (nmp->nm_so && adjsock) {
nfs_safedisconnect(nmp);
if (nmp->nm_sotype == SOCK_DGRAM)
while (nfs_connect(nmp, NULL)) {
printf("nfs_args: retrying connect\n");
(void) tsleep((caddr_t)&lbolt, PSOCK, "nfscon", 0);
}
}
strlcpy(nmp->nm_hostname, argp->hostname, sizeof(nmp->nm_hostname));
p = strchr(nmp->nm_hostname, ':');
if (p)
*p = '\0';
}
static const char *nfs_opts[] = { "from", "nfs_args",
"noatime", "noexec", "suiddir", "nosuid", "nosymfollow", "union",
"noclusterr", "noclusterw", "multilabel", "acls", "force", "update",
"async", "dumbtimer", "noconn", "nolockd", "intr", "rdirplus", "resvport",
"readdirsize", "soft", "hard", "mntudp", "tcp", "wsize", "rsize",
"retrans", "acregmin", "acregmax", "acdirmin", "acdirmax",
NULL };
The remaining part of nmount/omount/rootfs mount changes. I cannot sensibly split the conversion of the remaining three filesystems out from the root mounting changes, so in one go: cd9660: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() nfs(client): Convert to nmount (the simple way, mount_nfs(8) is still necessary). Add omount compat shims. Drop COMPAT_PRELITE2 mount arg compatibility. ffs: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() Remove vfs_omount() method, all filesystems are now converted. Remove MNTK_WANTRDWR, handling RO/RW conversions is a filesystem task, and they all do it now. Change rootmounting to use DEVFS trampoline: vfs_mount.c: Mount devfs on /. Devfs needs no 'from' so this is clean. symlink /dev to /. This makes it possible to lookup /dev/foo. Mount "real" root filesystem on /. Surgically move the devfs mountpoint from under the real root filesystem onto /dev in the real root filesystem. Remove now unnecessary getdiskbyname(). kern_init.c: Don't do devfs mounting and rootvnode assignment here, it was already handled by vfs_mount.c. Remove now unused bdevvp(), addaliasu() and addalias(). Put the few necessary lines in devfs where they belong. This eliminates the second-last source of bogo vnodes, leaving only the lemming-syncer. Remove rootdev variable, it doesn't give meaning in a global context and was not trustworth anyway. Correct information is provided by statfs(/).
2004-12-07 08:15:41 +00:00
1994-05-24 10:09:53 +00:00
/*
* VFS Operations.
*
* mount system call
* It seems a bit dumb to copyinstr() the host and path here and then
* bcopy() them in mountnfs(), but I wanted to detect errors before
* doing the sockargs() call because sockargs() allocates an mbuf and
* an error after that means that I have to release the mbuf.
*/
/* ARGSUSED */
1995-12-17 21:14:36 +00:00
static int
The remaining part of nmount/omount/rootfs mount changes. I cannot sensibly split the conversion of the remaining three filesystems out from the root mounting changes, so in one go: cd9660: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() nfs(client): Convert to nmount (the simple way, mount_nfs(8) is still necessary). Add omount compat shims. Drop COMPAT_PRELITE2 mount arg compatibility. ffs: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() Remove vfs_omount() method, all filesystems are now converted. Remove MNTK_WANTRDWR, handling RO/RW conversions is a filesystem task, and they all do it now. Change rootmounting to use DEVFS trampoline: vfs_mount.c: Mount devfs on /. Devfs needs no 'from' so this is clean. symlink /dev to /. This makes it possible to lookup /dev/foo. Mount "real" root filesystem on /. Surgically move the devfs mountpoint from under the real root filesystem onto /dev in the real root filesystem. Remove now unnecessary getdiskbyname(). kern_init.c: Don't do devfs mounting and rootvnode assignment here, it was already handled by vfs_mount.c. Remove now unused bdevvp(), addaliasu() and addalias(). Put the few necessary lines in devfs where they belong. This eliminates the second-last source of bogo vnodes, leaving only the lemming-syncer. Remove rootdev variable, it doesn't give meaning in a global context and was not trustworth anyway. Correct information is provided by statfs(/).
2004-12-07 08:15:41 +00:00
nfs_mount(struct mount *mp, struct thread *td)
1994-05-24 10:09:53 +00:00
{
struct nfs_args args = {
.version = NFS_ARGSVERSION,
.addr = NULL,
.addrlen = sizeof (struct sockaddr_in),
.sotype = SOCK_STREAM,
.proto = 0,
.fh = NULL,
.fhsize = 0,
.flags = NFSMNT_RESVPORT,
.wsize = NFS_WSIZE,
.rsize = NFS_RSIZE,
.readdirsize = NFS_READDIRSIZE,
.timeo = 10,
.retrans = NFS_RETRANS,
.maxgrouplist = NFS_MAXGRPS,
.readahead = NFS_DEFRAHEAD,
.wcommitsize = 0, /* was: NQ_DEFLEASE */
.deadthresh = NFS_MAXDEADTHRESH, /* was: NQ_DEADTHRESH */
.hostname = NULL,
/* args version 4 */
.acregmin = NFS_MINATTRTIMO,
.acregmax = NFS_MAXATTRTIMO,
.acdirmin = NFS_MINDIRATTRTIMO,
.acdirmax = NFS_MAXDIRATTRTIMO,
};
1994-05-24 10:09:53 +00:00
int error;
struct sockaddr *nam;
1994-05-24 10:09:53 +00:00
struct vnode *vp;
char hst[MNAMELEN];
size_t len;
u_char nfh[NFSX_V3FHMAX];
The remaining part of nmount/omount/rootfs mount changes. I cannot sensibly split the conversion of the remaining three filesystems out from the root mounting changes, so in one go: cd9660: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() nfs(client): Convert to nmount (the simple way, mount_nfs(8) is still necessary). Add omount compat shims. Drop COMPAT_PRELITE2 mount arg compatibility. ffs: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() Remove vfs_omount() method, all filesystems are now converted. Remove MNTK_WANTRDWR, handling RO/RW conversions is a filesystem task, and they all do it now. Change rootmounting to use DEVFS trampoline: vfs_mount.c: Mount devfs on /. Devfs needs no 'from' so this is clean. symlink /dev to /. This makes it possible to lookup /dev/foo. Mount "real" root filesystem on /. Surgically move the devfs mountpoint from under the real root filesystem onto /dev in the real root filesystem. Remove now unnecessary getdiskbyname(). kern_init.c: Don't do devfs mounting and rootvnode assignment here, it was already handled by vfs_mount.c. Remove now unused bdevvp(), addaliasu() and addalias(). Put the few necessary lines in devfs where they belong. This eliminates the second-last source of bogo vnodes, leaving only the lemming-syncer. Remove rootdev variable, it doesn't give meaning in a global context and was not trustworth anyway. Correct information is provided by statfs(/).
2004-12-07 08:15:41 +00:00
if (vfs_filteropt(mp->mnt_optnew, nfs_opts)) {
error = EINVAL;
goto out;
}
1994-05-24 10:09:53 +00:00
if ((mp->mnt_flag & (MNT_ROOTFS | MNT_UPDATE)) == MNT_ROOTFS) {
error = nfs_mountroot(mp, td);
goto out;
}
The remaining part of nmount/omount/rootfs mount changes. I cannot sensibly split the conversion of the remaining three filesystems out from the root mounting changes, so in one go: cd9660: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() nfs(client): Convert to nmount (the simple way, mount_nfs(8) is still necessary). Add omount compat shims. Drop COMPAT_PRELITE2 mount arg compatibility. ffs: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() Remove vfs_omount() method, all filesystems are now converted. Remove MNTK_WANTRDWR, handling RO/RW conversions is a filesystem task, and they all do it now. Change rootmounting to use DEVFS trampoline: vfs_mount.c: Mount devfs on /. Devfs needs no 'from' so this is clean. symlink /dev to /. This makes it possible to lookup /dev/foo. Mount "real" root filesystem on /. Surgically move the devfs mountpoint from under the real root filesystem onto /dev in the real root filesystem. Remove now unnecessary getdiskbyname(). kern_init.c: Don't do devfs mounting and rootvnode assignment here, it was already handled by vfs_mount.c. Remove now unused bdevvp(), addaliasu() and addalias(). Put the few necessary lines in devfs where they belong. This eliminates the second-last source of bogo vnodes, leaving only the lemming-syncer. Remove rootdev variable, it doesn't give meaning in a global context and was not trustworth anyway. Correct information is provided by statfs(/).
2004-12-07 08:15:41 +00:00
error = vfs_copyopt(mp->mnt_optnew, "nfs_args", &args, sizeof args);
if (error)
goto out;
The remaining part of nmount/omount/rootfs mount changes. I cannot sensibly split the conversion of the remaining three filesystems out from the root mounting changes, so in one go: cd9660: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() nfs(client): Convert to nmount (the simple way, mount_nfs(8) is still necessary). Add omount compat shims. Drop COMPAT_PRELITE2 mount arg compatibility. ffs: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() Remove vfs_omount() method, all filesystems are now converted. Remove MNTK_WANTRDWR, handling RO/RW conversions is a filesystem task, and they all do it now. Change rootmounting to use DEVFS trampoline: vfs_mount.c: Mount devfs on /. Devfs needs no 'from' so this is clean. symlink /dev to /. This makes it possible to lookup /dev/foo. Mount "real" root filesystem on /. Surgically move the devfs mountpoint from under the real root filesystem onto /dev in the real root filesystem. Remove now unnecessary getdiskbyname(). kern_init.c: Don't do devfs mounting and rootvnode assignment here, it was already handled by vfs_mount.c. Remove now unused bdevvp(), addaliasu() and addalias(). Put the few necessary lines in devfs where they belong. This eliminates the second-last source of bogo vnodes, leaving only the lemming-syncer. Remove rootdev variable, it doesn't give meaning in a global context and was not trustworth anyway. Correct information is provided by statfs(/).
2004-12-07 08:15:41 +00:00
if (args.version != NFS_ARGSVERSION) {
error = EPROGMISMATCH;
goto out;
}
if (mp->mnt_flag & MNT_UPDATE) {
struct nfsmount *nmp = VFSTONFS(mp);
if (nmp == NULL) {
error = EIO;
goto out;
}
/*
* When doing an update, we can't change from or to
* v3, switch lockd strategies or change cookie translation
*/
args.flags = (args.flags &
~(NFSMNT_NFSV3 | NFSMNT_NOLOCKD /*|NFSMNT_XLATECOOKIE*/)) |
(nmp->nm_flag &
(NFSMNT_NFSV3 | NFSMNT_NOLOCKD /*|NFSMNT_XLATECOOKIE*/));
The remaining part of nmount/omount/rootfs mount changes. I cannot sensibly split the conversion of the remaining three filesystems out from the root mounting changes, so in one go: cd9660: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() nfs(client): Convert to nmount (the simple way, mount_nfs(8) is still necessary). Add omount compat shims. Drop COMPAT_PRELITE2 mount arg compatibility. ffs: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() Remove vfs_omount() method, all filesystems are now converted. Remove MNTK_WANTRDWR, handling RO/RW conversions is a filesystem task, and they all do it now. Change rootmounting to use DEVFS trampoline: vfs_mount.c: Mount devfs on /. Devfs needs no 'from' so this is clean. symlink /dev to /. This makes it possible to lookup /dev/foo. Mount "real" root filesystem on /. Surgically move the devfs mountpoint from under the real root filesystem onto /dev in the real root filesystem. Remove now unnecessary getdiskbyname(). kern_init.c: Don't do devfs mounting and rootvnode assignment here, it was already handled by vfs_mount.c. Remove now unused bdevvp(), addaliasu() and addalias(). Put the few necessary lines in devfs where they belong. This eliminates the second-last source of bogo vnodes, leaving only the lemming-syncer. Remove rootdev variable, it doesn't give meaning in a global context and was not trustworth anyway. Correct information is provided by statfs(/).
2004-12-07 08:15:41 +00:00
nfs_decode_args(mp, nmp, &args);
goto out;
}
/*
* Make the nfs_ip_paranoia sysctl serve as the default connection
* or no-connection mode for those protocols that support
* no-connection mode (the flag will be cleared later for protocols
* that do not support no-connection mode). This will allow a client
* to receive replies from a different IP then the request was
* sent to. Note: default value for nfs_ip_paranoia is 1 (paranoid),
* not 0.
*/
if (nfs_ip_paranoia == 0)
args.flags |= NFSMNT_NOCONN;
if (args.fhsize < 0 || args.fhsize > NFSX_V3FHMAX) {
error = EINVAL;
goto out;
}
error = copyin((caddr_t)args.fh, (caddr_t)nfh, args.fhsize);
if (error)
goto out;
error = copyinstr(args.hostname, hst, MNAMELEN-1, &len);
if (error)
goto out;
1994-05-24 10:09:53 +00:00
bzero(&hst[len], MNAMELEN - len);
/* sockargs() call must be after above copyin() calls */
error = getsockaddr(&nam, (caddr_t)args.addr, args.addrlen);
if (error)
goto out;
args.fh = nfh;
2004-12-07 14:26:39 +00:00
error = mountnfs(&args, mp, nam, hst, &vp, td->td_ucred);
out:
if (!error) {
MNT_ILOCK(mp);
mp->mnt_kern_flag |= (MNTK_MPSAFE|MNTK_LOOKUP_SHARED);
MNT_IUNLOCK(mp);
}
1994-05-24 10:09:53 +00:00
return (error);
}
The remaining part of nmount/omount/rootfs mount changes. I cannot sensibly split the conversion of the remaining three filesystems out from the root mounting changes, so in one go: cd9660: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() nfs(client): Convert to nmount (the simple way, mount_nfs(8) is still necessary). Add omount compat shims. Drop COMPAT_PRELITE2 mount arg compatibility. ffs: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() Remove vfs_omount() method, all filesystems are now converted. Remove MNTK_WANTRDWR, handling RO/RW conversions is a filesystem task, and they all do it now. Change rootmounting to use DEVFS trampoline: vfs_mount.c: Mount devfs on /. Devfs needs no 'from' so this is clean. symlink /dev to /. This makes it possible to lookup /dev/foo. Mount "real" root filesystem on /. Surgically move the devfs mountpoint from under the real root filesystem onto /dev in the real root filesystem. Remove now unnecessary getdiskbyname(). kern_init.c: Don't do devfs mounting and rootvnode assignment here, it was already handled by vfs_mount.c. Remove now unused bdevvp(), addaliasu() and addalias(). Put the few necessary lines in devfs where they belong. This eliminates the second-last source of bogo vnodes, leaving only the lemming-syncer. Remove rootdev variable, it doesn't give meaning in a global context and was not trustworth anyway. Correct information is provided by statfs(/).
2004-12-07 08:15:41 +00:00
/*
* VFS Operations.
*
* mount system call
* It seems a bit dumb to copyinstr() the host and path here and then
* bcopy() them in mountnfs(), but I wanted to detect errors before
* doing the sockargs() call because sockargs() allocates an mbuf and
* an error after that means that I have to release the mbuf.
*/
/* ARGSUSED */
static int
nfs_cmount(struct mntarg *ma, void *data, int flags, struct thread *td)
{
int error;
struct nfs_args args;
error = copyin(data, &args, sizeof (struct nfs_args));
if (error)
return error;
The remaining part of nmount/omount/rootfs mount changes. I cannot sensibly split the conversion of the remaining three filesystems out from the root mounting changes, so in one go: cd9660: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() nfs(client): Convert to nmount (the simple way, mount_nfs(8) is still necessary). Add omount compat shims. Drop COMPAT_PRELITE2 mount arg compatibility. ffs: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() Remove vfs_omount() method, all filesystems are now converted. Remove MNTK_WANTRDWR, handling RO/RW conversions is a filesystem task, and they all do it now. Change rootmounting to use DEVFS trampoline: vfs_mount.c: Mount devfs on /. Devfs needs no 'from' so this is clean. symlink /dev to /. This makes it possible to lookup /dev/foo. Mount "real" root filesystem on /. Surgically move the devfs mountpoint from under the real root filesystem onto /dev in the real root filesystem. Remove now unnecessary getdiskbyname(). kern_init.c: Don't do devfs mounting and rootvnode assignment here, it was already handled by vfs_mount.c. Remove now unused bdevvp(), addaliasu() and addalias(). Put the few necessary lines in devfs where they belong. This eliminates the second-last source of bogo vnodes, leaving only the lemming-syncer. Remove rootdev variable, it doesn't give meaning in a global context and was not trustworth anyway. Correct information is provided by statfs(/).
2004-12-07 08:15:41 +00:00
ma = mount_arg(ma, "nfs_args", &args, sizeof args);
error = kernel_mount(ma, flags);
return (error);
}
1994-05-24 10:09:53 +00:00
/*
* Common code for mount and mountroot
*/
1995-12-17 21:14:36 +00:00
static int
mountnfs(struct nfs_args *argp, struct mount *mp, struct sockaddr *nam,
2004-12-07 14:26:39 +00:00
char *hst, struct vnode **vpp, struct ucred *cred)
1994-05-24 10:09:53 +00:00
{
struct nfsmount *nmp;
1994-05-24 10:09:53 +00:00
struct nfsnode *np;
int error;
struct vattr attrs;
1994-05-24 10:09:53 +00:00
if (mp->mnt_flag & MNT_UPDATE) {
nmp = VFSTONFS(mp);
printf("%s: MNT_UPDATE is no longer handled here\n", __func__);
FREE(nam, M_SONAME);
1994-05-24 10:09:53 +00:00
return (0);
} else {
nmp = uma_zalloc(nfsmount_zone, M_WAITOK);
1994-05-24 10:09:53 +00:00
bzero((caddr_t)nmp, sizeof (struct nfsmount));
TAILQ_INIT(&nmp->nm_bufq);
2007-10-16 10:54:55 +00:00
mp->mnt_data = nmp;
1994-05-24 10:09:53 +00:00
}
vfs_getnewfsid(mp);
1994-05-24 10:09:53 +00:00
nmp->nm_mountp = mp;
mtx_init(&nmp->nm_mtx, "NFSmount lock", NULL, MTX_DEF);
/*
* V2 can only handle 32 bit filesizes. A 4GB-1 limit may be too
* high, depending on whether we end up with negative offsets in
* the client or server somewhere. 2GB-1 may be safer.
*
* For V3, nfs_fsinfo will adjust this as necessary. Assume maximum
* that we can handle until we find out otherwise.
* XXX Our "safe" limit on the client is what we can store in our
* buffer cache using signed(!) block numbers.
*/
1998-06-01 11:07:16 +00:00
if ((argp->flags & NFSMNT_NFSV3) == 0)
nmp->nm_maxfilesize = 0xffffffffLL;
else
nmp->nm_maxfilesize = (u_int64_t)0x80000000 * DEV_BSIZE - 1;
1994-05-24 10:09:53 +00:00
nmp->nm_timeo = NFS_TIMEO;
nmp->nm_retry = NFS_RETRANS;
if ((argp->flags & NFSMNT_NFSV3) && argp->sotype == SOCK_STREAM) {
nmp->nm_wsize = nmp->nm_rsize = NFS_MAXDATA;
} else {
nmp->nm_wsize = NFS_WSIZE;
nmp->nm_rsize = NFS_RSIZE;
}
nmp->nm_wcommitsize = hibufspace / (desiredvnodes / 1000);
nmp->nm_readdirsize = NFS_READDIRSIZE;
1994-05-24 10:09:53 +00:00
nmp->nm_numgrps = NFS_MAXGRPS;
nmp->nm_readahead = NFS_DEFRAHEAD;
nmp->nm_deadthresh = NFS_MAXDEADTHRESH;
nmp->nm_tprintf_delay = nfs_tprintf_delay;
if (nmp->nm_tprintf_delay < 0)
nmp->nm_tprintf_delay = 0;
nmp->nm_tprintf_initial_delay = nfs_tprintf_initial_delay;
if (nmp->nm_tprintf_initial_delay < 0)
nmp->nm_tprintf_initial_delay = 0;
nmp->nm_fhsize = argp->fhsize;
bcopy((caddr_t)argp->fh, (caddr_t)nmp->nm_fh, argp->fhsize);
1994-05-24 10:09:53 +00:00
bcopy(hst, mp->mnt_stat.f_mntfromname, MNAMELEN);
nmp->nm_nam = nam;
/* Set up the sockets and per-host congestion */
nmp->nm_sotype = argp->sotype;
nmp->nm_soproto = argp->proto;
nmp->nm_rpcops = &nfs_rpcops;
1994-05-24 10:09:53 +00:00
The remaining part of nmount/omount/rootfs mount changes. I cannot sensibly split the conversion of the remaining three filesystems out from the root mounting changes, so in one go: cd9660: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() nfs(client): Convert to nmount (the simple way, mount_nfs(8) is still necessary). Add omount compat shims. Drop COMPAT_PRELITE2 mount arg compatibility. ffs: Convert to nmount. Add omount compat shims. Remove dedicated rootfs mounting code. Use vfs_mountedfrom() Rely on vfs_mount.c calling VFS_STATFS() Remove vfs_omount() method, all filesystems are now converted. Remove MNTK_WANTRDWR, handling RO/RW conversions is a filesystem task, and they all do it now. Change rootmounting to use DEVFS trampoline: vfs_mount.c: Mount devfs on /. Devfs needs no 'from' so this is clean. symlink /dev to /. This makes it possible to lookup /dev/foo. Mount "real" root filesystem on /. Surgically move the devfs mountpoint from under the real root filesystem onto /dev in the real root filesystem. Remove now unnecessary getdiskbyname(). kern_init.c: Don't do devfs mounting and rootvnode assignment here, it was already handled by vfs_mount.c. Remove now unused bdevvp(), addaliasu() and addalias(). Put the few necessary lines in devfs where they belong. This eliminates the second-last source of bogo vnodes, leaving only the lemming-syncer. Remove rootdev variable, it doesn't give meaning in a global context and was not trustworth anyway. Correct information is provided by statfs(/).
2004-12-07 08:15:41 +00:00
nfs_decode_args(mp, nmp, argp);
1994-05-24 10:09:53 +00:00
/*
* For Connection based sockets (TCP,...) defer the connect until
* the first request, in case the server is not responding.
*/
if (nmp->nm_sotype == SOCK_DGRAM &&
(error = nfs_connect(nmp, NULL)))
1994-05-24 10:09:53 +00:00
goto bad;
/*
* This is silly, but it has to be set so that vinifod() works.
* We do not want to do an nfs_statfs() here since we can get
* stuck on a dead server and we are holding a lock on the mount
* point.
*/
mtx_lock(&nmp->nm_mtx);
mp->mnt_stat.f_iosize = nfs_iosize(nmp);
mtx_unlock(&nmp->nm_mtx);
1994-05-24 10:09:53 +00:00
/*
* A reference count is needed on the nfsnode representing the
* remote root. If this object is not persistent, then backward
* traversals of the mount point (i.e. "..") will not work if
* the nfsnode gets flushed out of the cache. Ufs does not have
* this problem, because one can identify root inodes by their
* number == ROOTINO (2).
*/
error = nfs_nget(mp, (nfsfh_t *)nmp->nm_fh, nmp->nm_fhsize, &np, LK_EXCLUSIVE);
if (error)
1994-05-24 10:09:53 +00:00
goto bad;
*vpp = NFSTOV(np);
/*
* Get file attributes and transfer parameters for the
* mountpoint. This has the side effect of filling in
* (*vpp)->v_type with the correct value.
*/
if (argp->flags & NFSMNT_NFSV3)
nfs_fsinfo(nmp, *vpp, curthread->td_ucred, curthread);
else
VOP_GETATTR(*vpp, &attrs, curthread->td_ucred, curthread);
/*
* Lose the lock but keep the ref.
*/
VOP_UNLOCK(*vpp, 0);
1994-05-24 10:09:53 +00:00
return (0);
bad:
nfs_disconnect(nmp);
mtx_destroy(&nmp->nm_mtx);
uma_zfree(nfsmount_zone, nmp);
FREE(nam, M_SONAME);
1994-05-24 10:09:53 +00:00
return (error);
}
/*
* unmount system call
*/
1995-12-17 21:14:36 +00:00
static int
nfs_unmount(struct mount *mp, int mntflags, struct thread *td)
1994-05-24 10:09:53 +00:00
{
struct nfsmount *nmp;
1994-05-24 10:09:53 +00:00
int error, flags = 0;
if (mntflags & MNT_FORCE)
1994-05-24 10:09:53 +00:00
flags |= FORCECLOSE;
nmp = VFSTONFS(mp);
/*
* Goes something like this..
2002-05-16 21:28:32 +00:00
* - Call vflush() to clear out vnodes for this filesystem
1994-05-24 10:09:53 +00:00
* - Close the socket
* - Free up the data structures
*/
/* In the forced case, cancel any outstanding requests. */
if (flags & FORCECLOSE) {
error = nfs_nmcancelreqs(nmp);
if (error)
goto out;
}
/* We hold 1 extra ref on the root vnode; see comment in mountnfs(). */
error = vflush(mp, 1, flags, td);
if (error)
goto out;
1994-05-24 10:09:53 +00:00
/*
* We are now committed to the unmount.
*/
nfs_disconnect(nmp);
FREE(nmp->nm_nam, M_SONAME);
1994-05-24 10:09:53 +00:00
mtx_destroy(&nmp->nm_mtx);
uma_zfree(nfsmount_zone, nmp);
out:
return (error);
1994-05-24 10:09:53 +00:00
}
/*
* Return root of a filesystem
*/
1995-12-17 21:14:36 +00:00
static int
nfs_root(struct mount *mp, int flags, struct vnode **vpp, struct thread *td)
1994-05-24 10:09:53 +00:00
{
struct vnode *vp;
1994-05-24 10:09:53 +00:00
struct nfsmount *nmp;
struct nfsnode *np;
int error;
nmp = VFSTONFS(mp);
error = nfs_nget(mp, (nfsfh_t *)nmp->nm_fh, nmp->nm_fhsize, &np, flags);
if (error)
return error;
1994-05-24 10:09:53 +00:00
vp = NFSTOV(np);
/*
* Get transfer parameters and attributes for root vnode once.
*/
mtx_lock(&nmp->nm_mtx);
if ((nmp->nm_state & NFSSTA_GOTFSINFO) == 0 &&
(nmp->nm_flag & NFSMNT_NFSV3)) {
mtx_unlock(&nmp->nm_mtx);
nfs_fsinfo(nmp, vp, curthread->td_ucred, curthread);
} else
mtx_unlock(&nmp->nm_mtx);
if (vp->v_type == VNON)
vp->v_type = VDIR;
vp->v_vflag |= VV_ROOT;
1994-05-24 10:09:53 +00:00
*vpp = vp;
return (0);
}
/*
* Flush out the buffer cache
*/
/* ARGSUSED */
1995-12-17 21:14:36 +00:00
static int
nfs_sync(struct mount *mp, int waitfor, struct thread *td)
1994-05-24 10:09:53 +00:00
{
struct vnode *vp, *mvp;
1994-05-24 10:09:53 +00:00
int error, allerror = 0;
/*
* Force stale buffer cache information to be flushed.
*/
MNT_ILOCK(mp);
1994-05-24 10:09:53 +00:00
loop:
MNT_VNODE_FOREACH(vp, mp, mvp) {
VI_LOCK(vp);
MNT_IUNLOCK(mp);
/* XXX Racy bv_cnt check. */
if (VOP_ISLOCKED(vp) || vp->v_bufobj.bo_dirty.bv_cnt == 0 ||
waitfor == MNT_LAZY) {
VI_UNLOCK(vp);
MNT_ILOCK(mp);
1994-05-24 10:09:53 +00:00
continue;
}
if (vget(vp, LK_EXCLUSIVE | LK_INTERLOCK, td)) {
MNT_ILOCK(mp);
MNT_VNODE_FOREACH_ABORT_ILOCKED(mp, mvp);
1994-05-24 10:09:53 +00:00
goto loop;
}
error = VOP_FSYNC(vp, waitfor, td);
if (error)
1994-05-24 10:09:53 +00:00
allerror = error;
VOP_UNLOCK(vp, 0);
vrele(vp);
MNT_ILOCK(mp);
1994-05-24 10:09:53 +00:00
}
MNT_IUNLOCK(mp);
1994-05-24 10:09:53 +00:00
return (allerror);
}
static int
nfs_sysctl(struct mount *mp, fsctlop_t op, struct sysctl_req *req)
{
struct nfsmount *nmp = VFSTONFS(mp);
struct vfsquery vq;
int error;
bzero(&vq, sizeof(vq));
switch (op) {
#if 0
case VFS_CTL_NOLOCKS:
val = (nmp->nm_flag & NFSMNT_NOLOCKS) ? 1 : 0;
if (req->oldptr != NULL) {
error = SYSCTL_OUT(req, &val, sizeof(val));
if (error)
return (error);
}
if (req->newptr != NULL) {
error = SYSCTL_IN(req, &val, sizeof(val));
if (error)
return (error);
if (val)
nmp->nm_flag |= NFSMNT_NOLOCKS;
else
nmp->nm_flag &= ~NFSMNT_NOLOCKS;
}
break;
#endif
case VFS_CTL_QUERY:
mtx_lock(&nmp->nm_mtx);
if (nmp->nm_state & NFSSTA_TIMEO)
vq.vq_flags |= VQ_NOTRESP;
mtx_unlock(&nmp->nm_mtx);
#if 0
if (!(nmp->nm_flag & NFSMNT_NOLOCKS) &&
(nmp->nm_state & NFSSTA_LOCKTIMEO))
vq.vq_flags |= VQ_NOTRESPLOCK;
#endif
error = SYSCTL_OUT(req, &vq, sizeof(vq));
break;
case VFS_CTL_TIMEO:
if (req->oldptr != NULL) {
error = SYSCTL_OUT(req, &nmp->nm_tprintf_initial_delay,
sizeof(nmp->nm_tprintf_initial_delay));
if (error)
return (error);
}
if (req->newptr != NULL) {
error = vfs_suser(mp, req->td);
if (error)
return (error);
error = SYSCTL_IN(req, &nmp->nm_tprintf_initial_delay,
sizeof(nmp->nm_tprintf_initial_delay));
if (error)
return (error);
if (nmp->nm_tprintf_initial_delay < 0)
nmp->nm_tprintf_initial_delay = 0;
}
break;
default:
return (ENOTSUP);
}
return (0);
}