1
0
mirror of https://git.FreeBSD.org/src.git synced 2024-12-25 11:37:56 +00:00
freebsd/sys/kern/vfs_default.c
Alan Cox 4221e284a3 The VFS/BIO subsystem contained a number of hacks in order to optimize
piecemeal, middle-of-file writes for NFS.  These hacks have caused no
end of trouble, especially when combined with mmap().  I've removed
them.  Instead, NFS will issue a read-before-write to fully
instantiate the struct buf containing the write.  NFS does, however,
optimize piecemeal appends to files.  For most common file operations,
you will not notice the difference.  The sole remaining fragment in
the VFS/BIO system is b_dirtyoff/end, which NFS uses to avoid cache
coherency issues with read-merge-write style operations.  NFS also
optimizes the write-covers-entire-buffer case by avoiding the
read-before-write.  There is quite a bit of room for further
optimization in these areas.

The VM system marks pages fully-valid (AKA vm_page_t->valid =
VM_PAGE_BITS_ALL) in several places, most noteably in vm_fault.  This
is not correct operation.  The vm_pager_get_pages() code is now
responsible for marking VM pages all-valid.  A number of VM helper
routines have been added to aid in zeroing-out the invalid portions of
a VM page prior to the page being marked all-valid.  This operation is
necessary to properly support mmap().  The zeroing occurs most often
when dealing with file-EOF situations.  Several bugs have been fixed
in the NFS subsystem, including bits handling file and directory EOF
situations and buf->b_flags consistancy issues relating to clearing
B_ERROR & B_INVAL, and handling B_DONE.

getblk() and allocbuf() have been rewritten.  B_CACHE operation is now
formally defined in comments and more straightforward in
implementation.  B_CACHE for VMIO buffers is based on the validity of
the backing store.  B_CACHE for non-VMIO buffers is based simply on
whether the buffer is B_INVAL or not (B_CACHE set if B_INVAL clear,
and vise-versa).  biodone() is now responsible for setting B_CACHE
when a successful read completes.  B_CACHE is also set when a bdwrite()
is initiated and when a bwrite() is initiated.  VFS VOP_BWRITE
routines (there are only two - nfs_bwrite() and bwrite()) are now
expected to set B_CACHE.  This means that bowrite() and bawrite() also
set B_CACHE indirectly.

There are a number of places in the code which were previously using
buf->b_bufsize (which is DEV_BSIZE aligned) when they should have
been using buf->b_bcount.  These have been fixed.  getblk() now clears
B_DONE on return because the rest of the system is so bad about
dealing with B_DONE.

Major fixes to NFS/TCP have been made.  A server-side bug could cause
requests to be lost by the server due to nfs_realign() overwriting
other rpc's in the same TCP mbuf chain.  The server's kernel must be
recompiled to get the benefit of the fixes.

Submitted by:	Matthew Dillon <dillon@apollo.backplane.com>
1999-05-02 23:57:16 +00:00

491 lines
13 KiB
C

/*
* Copyright (c) 1989, 1993
* The Regents of the University of California. All rights reserved.
*
* This code is derived from software contributed
* to Berkeley by John Heidemann of the UCLA Ficus project.
*
* Source: * @(#)i405_init.c 2.10 92/04/27 UCLA Ficus project
*
* 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. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by the University of
* California, Berkeley and its contributors.
* 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.
*
*/
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/buf.h>
#include <sys/kernel.h>
#include <sys/lock.h>
#include <sys/malloc.h>
#include <sys/unistd.h>
#include <sys/vnode.h>
#include <sys/poll.h>
static int vop_nostrategy __P((struct vop_strategy_args *));
/*
* This vnode table stores what we want to do if the filesystem doesn't
* implement a particular VOP.
*
* If there is no specific entry here, we will return EOPNOTSUPP.
*
*/
vop_t **default_vnodeop_p;
static struct vnodeopv_entry_desc default_vnodeop_entries[] = {
{ &vop_default_desc, (vop_t *) vop_eopnotsupp },
{ &vop_abortop_desc, (vop_t *) vop_null },
{ &vop_advlock_desc, (vop_t *) vop_einval },
{ &vop_bwrite_desc, (vop_t *) vop_stdbwrite },
{ &vop_close_desc, (vop_t *) vop_null },
{ &vop_fsync_desc, (vop_t *) vop_null },
{ &vop_ioctl_desc, (vop_t *) vop_enotty },
{ &vop_islocked_desc, (vop_t *) vop_noislocked },
{ &vop_lease_desc, (vop_t *) vop_null },
{ &vop_lock_desc, (vop_t *) vop_nolock },
{ &vop_mmap_desc, (vop_t *) vop_einval },
{ &vop_open_desc, (vop_t *) vop_null },
{ &vop_pathconf_desc, (vop_t *) vop_einval },
{ &vop_poll_desc, (vop_t *) vop_nopoll },
{ &vop_readlink_desc, (vop_t *) vop_einval },
{ &vop_reallocblks_desc, (vop_t *) vop_eopnotsupp },
{ &vop_revoke_desc, (vop_t *) vop_revoke },
{ &vop_strategy_desc, (vop_t *) vop_nostrategy },
{ &vop_unlock_desc, (vop_t *) vop_nounlock },
{ NULL, NULL }
};
static struct vnodeopv_desc default_vnodeop_opv_desc =
{ &default_vnodeop_p, default_vnodeop_entries };
VNODEOP_SET(default_vnodeop_opv_desc);
int
vop_eopnotsupp(struct vop_generic_args *ap)
{
/*
printf("vop_notsupp[%s]\n", ap->a_desc->vdesc_name);
*/
return (EOPNOTSUPP);
}
int
vop_ebadf(struct vop_generic_args *ap)
{
return (EBADF);
}
int
vop_enotty(struct vop_generic_args *ap)
{
return (ENOTTY);
}
int
vop_einval(struct vop_generic_args *ap)
{
return (EINVAL);
}
int
vop_null(struct vop_generic_args *ap)
{
return (0);
}
int
vop_defaultop(struct vop_generic_args *ap)
{
return (VOCALL(default_vnodeop_p, ap->a_desc->vdesc_offset, ap));
}
int
vop_panic(struct vop_generic_args *ap)
{
panic("illegal vnode op called");
}
/*
* vop_nostrategy:
*
* Strategy routine for VFS devices that have none.
*
* B_ERROR and B_INVAL must be cleared prior to calling any strategy
* routine. Typically this is done for a B_READ strategy call. Typically
* B_INVAL is assumed to already be clear prior to a write and should not
* be cleared manually unless you just made the buffer invalid. B_ERROR
* should be cleared either way.
*/
static int
vop_nostrategy (struct vop_strategy_args *ap)
{
printf("No strategy for buffer at %p\n", ap->a_bp);
vprint("", ap->a_vp);
vprint("", ap->a_bp->b_vp);
ap->a_bp->b_flags |= B_ERROR;
ap->a_bp->b_error = EOPNOTSUPP;
biodone(ap->a_bp);
return (EOPNOTSUPP);
}
int
vop_stdpathconf(ap)
struct vop_pathconf_args /* {
struct vnode *a_vp;
int a_name;
int *a_retval;
} */ *ap;
{
switch (ap->a_name) {
case _PC_LINK_MAX:
*ap->a_retval = LINK_MAX;
return (0);
case _PC_MAX_CANON:
*ap->a_retval = MAX_CANON;
return (0);
case _PC_MAX_INPUT:
*ap->a_retval = MAX_INPUT;
return (0);
case _PC_PIPE_BUF:
*ap->a_retval = PIPE_BUF;
return (0);
case _PC_CHOWN_RESTRICTED:
*ap->a_retval = 1;
return (0);
case _PC_VDISABLE:
*ap->a_retval = _POSIX_VDISABLE;
return (0);
default:
return (EINVAL);
}
/* NOTREACHED */
}
/*
* Standard lock, unlock and islocked functions.
*
* These depend on the lock structure being the first element in the
* inode, ie: vp->v_data points to the the lock!
*/
int
vop_stdlock(ap)
struct vop_lock_args /* {
struct vnode *a_vp;
int a_flags;
struct proc *a_p;
} */ *ap;
{
struct lock *l;
if ((l = (struct lock *)ap->a_vp->v_data) == NULL) {
if (ap->a_flags & LK_INTERLOCK)
simple_unlock(&ap->a_vp->v_interlock);
return 0;
}
#ifndef DEBUG_LOCKS
return (lockmgr(l, ap->a_flags, &ap->a_vp->v_interlock, ap->a_p));
#else
return (debuglockmgr(l, ap->a_flags, &ap->a_vp->v_interlock, ap->a_p,
"vop_stdlock", ap->a_vp->filename, ap->a_vp->line));
#endif
}
int
vop_stdunlock(ap)
struct vop_unlock_args /* {
struct vnode *a_vp;
int a_flags;
struct proc *a_p;
} */ *ap;
{
struct lock *l;
if ((l = (struct lock *)ap->a_vp->v_data) == NULL) {
if (ap->a_flags & LK_INTERLOCK)
simple_unlock(&ap->a_vp->v_interlock);
return 0;
}
return (lockmgr(l, ap->a_flags | LK_RELEASE, &ap->a_vp->v_interlock,
ap->a_p));
}
int
vop_stdislocked(ap)
struct vop_islocked_args /* {
struct vnode *a_vp;
} */ *ap;
{
struct lock *l;
if ((l = (struct lock *)ap->a_vp->v_data) == NULL)
return 0;
return (lockstatus(l));
}
/*
* Return true for select/poll.
*/
int
vop_nopoll(ap)
struct vop_poll_args /* {
struct vnode *a_vp;
int a_events;
struct ucred *a_cred;
struct proc *a_p;
} */ *ap;
{
/*
* Return true for read/write. If the user asked for something
* special, return POLLNVAL, so that clients have a way of
* determining reliably whether or not the extended
* functionality is present without hard-coding knowledge
* of specific filesystem implementations.
*/
if (ap->a_events & ~POLLSTANDARD)
return (POLLNVAL);
return (ap->a_events & (POLLIN | POLLOUT | POLLRDNORM | POLLWRNORM));
}
/*
* Implement poll for local filesystems that support it.
*/
int
vop_stdpoll(ap)
struct vop_poll_args /* {
struct vnode *a_vp;
int a_events;
struct ucred *a_cred;
struct proc *a_p;
} */ *ap;
{
if ((ap->a_events & ~POLLSTANDARD) == 0)
return (ap->a_events & (POLLRDNORM|POLLWRNORM));
return (vn_pollrecord(ap->a_vp, ap->a_p, ap->a_events));
}
int
vop_stdbwrite(ap)
struct vop_bwrite_args *ap;
{
return (bwrite(ap->a_bp));
}
/*
* Stubs to use when there is no locking to be done on the underlying object.
* A minimal shared lock is necessary to ensure that the underlying object
* is not revoked while an operation is in progress. So, an active shared
* count is maintained in an auxillary vnode lock structure.
*/
int
vop_sharedlock(ap)
struct vop_lock_args /* {
struct vnode *a_vp;
int a_flags;
struct proc *a_p;
} */ *ap;
{
/*
* This code cannot be used until all the non-locking filesystems
* (notably NFS) are converted to properly lock and release nodes.
* Also, certain vnode operations change the locking state within
* the operation (create, mknod, remove, link, rename, mkdir, rmdir,
* and symlink). Ideally these operations should not change the
* lock state, but should be changed to let the caller of the
* function unlock them. Otherwise all intermediate vnode layers
* (such as union, umapfs, etc) must catch these functions to do
* the necessary locking at their layer. Note that the inactive
* and lookup operations also change their lock state, but this
* cannot be avoided, so these two operations will always need
* to be handled in intermediate layers.
*/
struct vnode *vp = ap->a_vp;
int vnflags, flags = ap->a_flags;
if (vp->v_vnlock == NULL) {
if ((flags & LK_TYPE_MASK) == LK_DRAIN)
return (0);
MALLOC(vp->v_vnlock, struct lock *, sizeof(struct lock),
M_VNODE, M_WAITOK);
lockinit(vp->v_vnlock, PVFS, "vnlock", 0, LK_NOPAUSE);
}
switch (flags & LK_TYPE_MASK) {
case LK_DRAIN:
vnflags = LK_DRAIN;
break;
case LK_EXCLUSIVE:
#ifdef DEBUG_VFS_LOCKS
/*
* Normally, we use shared locks here, but that confuses
* the locking assertions.
*/
vnflags = LK_EXCLUSIVE;
break;
#endif
case LK_SHARED:
vnflags = LK_SHARED;
break;
case LK_UPGRADE:
case LK_EXCLUPGRADE:
case LK_DOWNGRADE:
return (0);
case LK_RELEASE:
default:
panic("vop_sharedlock: bad operation %d", flags & LK_TYPE_MASK);
}
if (flags & LK_INTERLOCK)
vnflags |= LK_INTERLOCK;
#ifndef DEBUG_LOCKS
return (lockmgr(vp->v_vnlock, vnflags, &vp->v_interlock, ap->a_p));
#else
return (debuglockmgr(vp->v_vnlock, vnflags, &vp->v_interlock, ap->a_p,
"vop_sharedlock", vp->filename, vp->line));
#endif
}
/*
* Stubs to use when there is no locking to be done on the underlying object.
* A minimal shared lock is necessary to ensure that the underlying object
* is not revoked while an operation is in progress. So, an active shared
* count is maintained in an auxillary vnode lock structure.
*/
int
vop_nolock(ap)
struct vop_lock_args /* {
struct vnode *a_vp;
int a_flags;
struct proc *a_p;
} */ *ap;
{
#ifdef notyet
/*
* This code cannot be used until all the non-locking filesystems
* (notably NFS) are converted to properly lock and release nodes.
* Also, certain vnode operations change the locking state within
* the operation (create, mknod, remove, link, rename, mkdir, rmdir,
* and symlink). Ideally these operations should not change the
* lock state, but should be changed to let the caller of the
* function unlock them. Otherwise all intermediate vnode layers
* (such as union, umapfs, etc) must catch these functions to do
* the necessary locking at their layer. Note that the inactive
* and lookup operations also change their lock state, but this
* cannot be avoided, so these two operations will always need
* to be handled in intermediate layers.
*/
struct vnode *vp = ap->a_vp;
int vnflags, flags = ap->a_flags;
if (vp->v_vnlock == NULL) {
if ((flags & LK_TYPE_MASK) == LK_DRAIN)
return (0);
MALLOC(vp->v_vnlock, struct lock *, sizeof(struct lock),
M_VNODE, M_WAITOK);
lockinit(vp->v_vnlock, PVFS, "vnlock", 0, LK_NOPAUSE);
}
switch (flags & LK_TYPE_MASK) {
case LK_DRAIN:
vnflags = LK_DRAIN;
break;
case LK_EXCLUSIVE:
case LK_SHARED:
vnflags = LK_SHARED;
break;
case LK_UPGRADE:
case LK_EXCLUPGRADE:
case LK_DOWNGRADE:
return (0);
case LK_RELEASE:
default:
panic("vop_nolock: bad operation %d", flags & LK_TYPE_MASK);
}
if (flags & LK_INTERLOCK)
vnflags |= LK_INTERLOCK;
return(lockmgr(vp->v_vnlock, vnflags, &vp->v_interlock, ap->a_p));
#else /* for now */
/*
* Since we are not using the lock manager, we must clear
* the interlock here.
*/
if (ap->a_flags & LK_INTERLOCK)
simple_unlock(&ap->a_vp->v_interlock);
return (0);
#endif
}
/*
* Do the inverse of vop_nolock, handling the interlock in a compatible way.
*/
int
vop_nounlock(ap)
struct vop_unlock_args /* {
struct vnode *a_vp;
int a_flags;
struct proc *a_p;
} */ *ap;
{
struct vnode *vp = ap->a_vp;
if (vp->v_vnlock == NULL) {
if (ap->a_flags & LK_INTERLOCK)
simple_unlock(&ap->a_vp->v_interlock);
return (0);
}
return (lockmgr(vp->v_vnlock, LK_RELEASE | ap->a_flags,
&ap->a_vp->v_interlock, ap->a_p));
}
/*
* Return whether or not the node is in use.
*/
int
vop_noislocked(ap)
struct vop_islocked_args /* {
struct vnode *a_vp;
} */ *ap;
{
struct vnode *vp = ap->a_vp;
if (vp->v_vnlock == NULL)
return (0);
return (lockstatus(vp->v_vnlock));
}