2001-09-23 02:44:08 +04:00
|
|
|
/* $NetBSD: filecore_vnops.c,v 1.15 2001/09/22 22:48:17 chs Exp $ */
|
1998-08-14 07:26:12 +04:00
|
|
|
|
|
|
|
/*-
|
|
|
|
* Copyright (c) 1998 Andrew McMurry
|
|
|
|
* Copyright (c) 1994 The Regents of the University of California.
|
|
|
|
* All rights reserved.
|
|
|
|
*
|
|
|
|
* 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.
|
|
|
|
*
|
1998-09-18 08:17:44 +04:00
|
|
|
* filecore_vnops.c 1.2 1998/8/18
|
1998-08-14 07:26:12 +04:00
|
|
|
*/
|
|
|
|
|
|
|
|
#include <sys/param.h>
|
|
|
|
#include <sys/systm.h>
|
|
|
|
#include <sys/namei.h>
|
|
|
|
#include <sys/resourcevar.h>
|
|
|
|
#include <sys/kernel.h>
|
|
|
|
#include <sys/file.h>
|
|
|
|
#include <sys/stat.h>
|
|
|
|
#include <sys/buf.h>
|
|
|
|
#include <sys/proc.h>
|
|
|
|
#include <sys/conf.h>
|
|
|
|
#include <sys/mount.h>
|
|
|
|
#include <sys/vnode.h>
|
|
|
|
#include <sys/malloc.h>
|
|
|
|
#include <sys/dirent.h>
|
|
|
|
|
|
|
|
#include <miscfs/genfs/genfs.h>
|
|
|
|
#include <miscfs/specfs/specdev.h>
|
|
|
|
|
|
|
|
#include <filecorefs/filecore.h>
|
|
|
|
#include <filecorefs/filecore_extern.h>
|
|
|
|
#include <filecorefs/filecore_node.h>
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Check mode permission on inode pointer. Mode is READ, WRITE or EXEC.
|
|
|
|
* The mode is shifted to select the owner/group/other fields. The
|
|
|
|
* super user is granted all permissions.
|
|
|
|
*/
|
|
|
|
int
|
|
|
|
filecore_access(v)
|
|
|
|
void *v;
|
|
|
|
{
|
|
|
|
struct vop_access_args /* {
|
|
|
|
struct vnode *a_vp;
|
|
|
|
int a_mode;
|
|
|
|
struct ucred *a_cred;
|
|
|
|
struct proc *a_p;
|
|
|
|
} */ *ap = v;
|
|
|
|
struct vnode *vp = ap->a_vp;
|
|
|
|
struct filecore_node *ip = VTOI(vp);
|
|
|
|
struct filecore_mnt *fcmp = ip->i_mnt;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Disallow write attempts unless the file is a socket,
|
|
|
|
* fifo, or a block or character device resident on the
|
|
|
|
* file system.
|
|
|
|
*/
|
|
|
|
if (ap->a_mode & VWRITE) {
|
|
|
|
switch (vp->v_type) {
|
|
|
|
case VDIR:
|
|
|
|
case VLNK:
|
|
|
|
case VREG:
|
|
|
|
return (EROFS);
|
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return (vaccess(vp->v_type, filecore_mode(ip),
|
|
|
|
fcmp->fc_uid, fcmp->fc_gid, ap->a_mode, ap->a_cred));
|
|
|
|
}
|
|
|
|
|
|
|
|
int
|
|
|
|
filecore_getattr(v)
|
|
|
|
void *v;
|
|
|
|
{
|
|
|
|
struct vop_getattr_args /* {
|
|
|
|
struct vnode *a_vp;
|
|
|
|
struct vattr *a_vap;
|
|
|
|
struct ucred *a_cred;
|
|
|
|
struct proc *a_p;
|
|
|
|
} */ *ap = v;
|
|
|
|
struct vnode *vp = ap->a_vp;
|
|
|
|
struct filecore_node *ip = VTOI(vp);
|
1998-08-14 22:04:04 +04:00
|
|
|
struct vattr *vap = ap->a_vap;
|
1998-08-14 07:26:12 +04:00
|
|
|
struct filecore_mnt *fcmp = ip->i_mnt;
|
|
|
|
|
|
|
|
vap->va_fsid = ip->i_dev;
|
|
|
|
vap->va_fileid = ip->i_number;
|
|
|
|
|
|
|
|
vap->va_mode = filecore_mode(ip);
|
|
|
|
vap->va_nlink = 1;
|
|
|
|
vap->va_uid = fcmp->fc_uid;
|
|
|
|
vap->va_gid = fcmp->fc_gid;
|
|
|
|
vap->va_atime = filecore_time(ip);
|
|
|
|
vap->va_mtime = filecore_time(ip);
|
|
|
|
vap->va_ctime = filecore_time(ip);
|
|
|
|
vap->va_rdev = 0; /* We don't support specials */
|
|
|
|
|
|
|
|
vap->va_size = (u_quad_t) ip->i_size;
|
|
|
|
vap->va_flags = 0;
|
1998-08-14 22:04:04 +04:00
|
|
|
vap->va_gen = 1;
|
1998-08-14 07:26:12 +04:00
|
|
|
vap->va_blocksize = fcmp->blksize;
|
|
|
|
vap->va_bytes = vap->va_size;
|
|
|
|
vap->va_type = vp->v_type;
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Vnode op for reading.
|
|
|
|
*/
|
|
|
|
int
|
|
|
|
filecore_read(v)
|
|
|
|
void *v;
|
|
|
|
{
|
|
|
|
struct vop_read_args /* {
|
|
|
|
struct vnode *a_vp;
|
|
|
|
struct uio *a_uio;
|
|
|
|
int a_ioflag;
|
|
|
|
struct ucred *a_cred;
|
|
|
|
} */ *ap = v;
|
|
|
|
struct vnode *vp = ap->a_vp;
|
1998-08-14 22:04:04 +04:00
|
|
|
struct uio *uio = ap->a_uio;
|
|
|
|
struct filecore_node *ip = VTOI(vp);
|
|
|
|
struct filecore_mnt *fcmp;
|
1998-08-14 07:26:12 +04:00
|
|
|
struct buf *bp;
|
|
|
|
daddr_t lbn, rablock;
|
|
|
|
off_t diff;
|
a whole bunch of changes to improve performance and robustness under load:
- remove special treatment of pager_map mappings in pmaps. this is
required now, since I've removed the globals that expose the address range.
pager_map now uses pmap_kenter_pa() instead of pmap_enter(), so there's
no longer any need to special-case it.
- eliminate struct uvm_vnode by moving its fields into struct vnode.
- rewrite the pageout path. the pager is now responsible for handling the
high-level requests instead of only getting control after a bunch of work
has already been done on its behalf. this will allow us to UBCify LFS,
which needs tighter control over its pages than other filesystems do.
writing a page to disk no longer requires making it read-only, which
allows us to write wired pages without causing all kinds of havoc.
- use a new PG_PAGEOUT flag to indicate that a page should be freed
on behalf of the pagedaemon when it's unlocked. this flag is very similar
to PG_RELEASED, but unlike PG_RELEASED, PG_PAGEOUT can be cleared if the
pageout fails due to eg. an indirect-block buffer being locked.
this allows us to remove the "version" field from struct vm_page,
and together with shrinking "loan_count" from 32 bits to 16,
struct vm_page is now 4 bytes smaller.
- no longer use PG_RELEASED for swap-backed pages. if the page is busy
because it's being paged out, we can't release the swap slot to be
reallocated until that write is complete, but unlike with vnodes we
don't keep a count of in-progress writes so there's no good way to
know when the write is done. instead, when we need to free a busy
swap-backed page, just sleep until we can get it busy ourselves.
- implement a fast-path for extending writes which allows us to avoid
zeroing new pages. this substantially reduces cpu usage.
- encapsulate the data used by the genfs code in a struct genfs_node,
which must be the first element of the filesystem-specific vnode data
for filesystems which use genfs_{get,put}pages().
- eliminate many of the UVM pagerops, since they aren't needed anymore
now that the pager "put" operation is a higher-level operation.
- enhance the genfs code to allow NFS to use the genfs_{get,put}pages
instead of a modified copy.
- clean up struct vnode by removing all the fields that used to be used by
the vfs_cluster.c code (which we don't use anymore with UBC).
- remove kmem_object and mb_object since they were useless.
instead of allocating pages to these objects, we now just allocate
pages with no object. such pages are mapped in the kernel until they
are freed, so we can use the mapping to find the page to free it.
this allows us to remove splvm() protection in several places.
The sum of all these changes improves write throughput on my
decstation 5000/200 to within 1% of the rate of NetBSD 1.5
and reduces the elapsed time for "make release" of a NetBSD 1.5
source tree on my 128MB pc to 10% less than a 1.5 kernel took.
2001-09-16 00:36:31 +04:00
|
|
|
int error = 0;
|
1998-08-14 07:26:12 +04:00
|
|
|
long size, n, on;
|
|
|
|
|
|
|
|
if (uio->uio_resid == 0)
|
|
|
|
return (0);
|
|
|
|
if (uio->uio_offset < 0)
|
|
|
|
return (EINVAL);
|
2001-09-23 02:44:08 +04:00
|
|
|
if (uio->uio_offset >= ip->i_size)
|
|
|
|
return (0);
|
1998-08-14 07:26:12 +04:00
|
|
|
ip->i_flag |= IN_ACCESS;
|
|
|
|
fcmp = ip->i_mnt;
|
2000-11-27 11:39:39 +03:00
|
|
|
|
|
|
|
if (vp->v_type == VREG) {
|
|
|
|
error = 0;
|
|
|
|
while (uio->uio_resid > 0) {
|
|
|
|
void *win;
|
2001-02-27 07:37:44 +03:00
|
|
|
vsize_t bytelen = MIN(ip->i_size - uio->uio_offset,
|
2000-11-27 11:39:39 +03:00
|
|
|
uio->uio_resid);
|
|
|
|
|
|
|
|
if (bytelen == 0) {
|
|
|
|
break;
|
|
|
|
}
|
a whole bunch of changes to improve performance and robustness under load:
- remove special treatment of pager_map mappings in pmaps. this is
required now, since I've removed the globals that expose the address range.
pager_map now uses pmap_kenter_pa() instead of pmap_enter(), so there's
no longer any need to special-case it.
- eliminate struct uvm_vnode by moving its fields into struct vnode.
- rewrite the pageout path. the pager is now responsible for handling the
high-level requests instead of only getting control after a bunch of work
has already been done on its behalf. this will allow us to UBCify LFS,
which needs tighter control over its pages than other filesystems do.
writing a page to disk no longer requires making it read-only, which
allows us to write wired pages without causing all kinds of havoc.
- use a new PG_PAGEOUT flag to indicate that a page should be freed
on behalf of the pagedaemon when it's unlocked. this flag is very similar
to PG_RELEASED, but unlike PG_RELEASED, PG_PAGEOUT can be cleared if the
pageout fails due to eg. an indirect-block buffer being locked.
this allows us to remove the "version" field from struct vm_page,
and together with shrinking "loan_count" from 32 bits to 16,
struct vm_page is now 4 bytes smaller.
- no longer use PG_RELEASED for swap-backed pages. if the page is busy
because it's being paged out, we can't release the swap slot to be
reallocated until that write is complete, but unlike with vnodes we
don't keep a count of in-progress writes so there's no good way to
know when the write is done. instead, when we need to free a busy
swap-backed page, just sleep until we can get it busy ourselves.
- implement a fast-path for extending writes which allows us to avoid
zeroing new pages. this substantially reduces cpu usage.
- encapsulate the data used by the genfs code in a struct genfs_node,
which must be the first element of the filesystem-specific vnode data
for filesystems which use genfs_{get,put}pages().
- eliminate many of the UVM pagerops, since they aren't needed anymore
now that the pager "put" operation is a higher-level operation.
- enhance the genfs code to allow NFS to use the genfs_{get,put}pages
instead of a modified copy.
- clean up struct vnode by removing all the fields that used to be used by
the vfs_cluster.c code (which we don't use anymore with UBC).
- remove kmem_object and mb_object since they were useless.
instead of allocating pages to these objects, we now just allocate
pages with no object. such pages are mapped in the kernel until they
are freed, so we can use the mapping to find the page to free it.
this allows us to remove splvm() protection in several places.
The sum of all these changes improves write throughput on my
decstation 5000/200 to within 1% of the rate of NetBSD 1.5
and reduces the elapsed time for "make release" of a NetBSD 1.5
source tree on my 128MB pc to 10% less than a 1.5 kernel took.
2001-09-16 00:36:31 +04:00
|
|
|
win = ubc_alloc(&vp->v_uobj, uio->uio_offset,
|
2000-11-27 11:39:39 +03:00
|
|
|
&bytelen, UBC_READ);
|
|
|
|
error = uiomove(win, bytelen, uio);
|
|
|
|
ubc_release(win, 0);
|
|
|
|
if (error) {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
1998-08-14 07:26:12 +04:00
|
|
|
do {
|
|
|
|
lbn = lblkno(fcmp, uio->uio_offset);
|
|
|
|
on = blkoff(fcmp, uio->uio_offset);
|
2001-02-27 07:37:44 +03:00
|
|
|
n = MIN(blksize(fcmp, ip, lbn) - on, uio->uio_resid);
|
1998-08-14 07:26:12 +04:00
|
|
|
diff = (off_t)ip->i_size - uio->uio_offset;
|
|
|
|
if (diff <= 0)
|
|
|
|
return (0);
|
|
|
|
if (diff < n)
|
|
|
|
n = diff;
|
|
|
|
size = blksize(fcmp, ip, lbn);
|
|
|
|
rablock = lbn + 1;
|
|
|
|
if (ip->i_dirent.attr & FILECORE_ATTR_DIR) {
|
|
|
|
error = filecore_dbread(ip, &bp);
|
|
|
|
on = uio->uio_offset;
|
2001-02-27 07:37:44 +03:00
|
|
|
n = MIN(FILECORE_DIR_SIZE - on, uio->uio_resid);
|
1998-08-14 07:26:12 +04:00
|
|
|
size = FILECORE_DIR_SIZE;
|
|
|
|
} else {
|
a whole bunch of changes to improve performance and robustness under load:
- remove special treatment of pager_map mappings in pmaps. this is
required now, since I've removed the globals that expose the address range.
pager_map now uses pmap_kenter_pa() instead of pmap_enter(), so there's
no longer any need to special-case it.
- eliminate struct uvm_vnode by moving its fields into struct vnode.
- rewrite the pageout path. the pager is now responsible for handling the
high-level requests instead of only getting control after a bunch of work
has already been done on its behalf. this will allow us to UBCify LFS,
which needs tighter control over its pages than other filesystems do.
writing a page to disk no longer requires making it read-only, which
allows us to write wired pages without causing all kinds of havoc.
- use a new PG_PAGEOUT flag to indicate that a page should be freed
on behalf of the pagedaemon when it's unlocked. this flag is very similar
to PG_RELEASED, but unlike PG_RELEASED, PG_PAGEOUT can be cleared if the
pageout fails due to eg. an indirect-block buffer being locked.
this allows us to remove the "version" field from struct vm_page,
and together with shrinking "loan_count" from 32 bits to 16,
struct vm_page is now 4 bytes smaller.
- no longer use PG_RELEASED for swap-backed pages. if the page is busy
because it's being paged out, we can't release the swap slot to be
reallocated until that write is complete, but unlike with vnodes we
don't keep a count of in-progress writes so there's no good way to
know when the write is done. instead, when we need to free a busy
swap-backed page, just sleep until we can get it busy ourselves.
- implement a fast-path for extending writes which allows us to avoid
zeroing new pages. this substantially reduces cpu usage.
- encapsulate the data used by the genfs code in a struct genfs_node,
which must be the first element of the filesystem-specific vnode data
for filesystems which use genfs_{get,put}pages().
- eliminate many of the UVM pagerops, since they aren't needed anymore
now that the pager "put" operation is a higher-level operation.
- enhance the genfs code to allow NFS to use the genfs_{get,put}pages
instead of a modified copy.
- clean up struct vnode by removing all the fields that used to be used by
the vfs_cluster.c code (which we don't use anymore with UBC).
- remove kmem_object and mb_object since they were useless.
instead of allocating pages to these objects, we now just allocate
pages with no object. such pages are mapped in the kernel until they
are freed, so we can use the mapping to find the page to free it.
this allows us to remove splvm() protection in several places.
The sum of all these changes improves write throughput on my
decstation 5000/200 to within 1% of the rate of NetBSD 1.5
and reduces the elapsed time for "make release" of a NetBSD 1.5
source tree on my 128MB pc to 10% less than a 1.5 kernel took.
2001-09-16 00:36:31 +04:00
|
|
|
error = bread(vp, lbn, size, NOCRED, &bp);
|
1998-08-14 07:26:12 +04:00
|
|
|
#ifdef FILECORE_DEBUG_BR
|
a whole bunch of changes to improve performance and robustness under load:
- remove special treatment of pager_map mappings in pmaps. this is
required now, since I've removed the globals that expose the address range.
pager_map now uses pmap_kenter_pa() instead of pmap_enter(), so there's
no longer any need to special-case it.
- eliminate struct uvm_vnode by moving its fields into struct vnode.
- rewrite the pageout path. the pager is now responsible for handling the
high-level requests instead of only getting control after a bunch of work
has already been done on its behalf. this will allow us to UBCify LFS,
which needs tighter control over its pages than other filesystems do.
writing a page to disk no longer requires making it read-only, which
allows us to write wired pages without causing all kinds of havoc.
- use a new PG_PAGEOUT flag to indicate that a page should be freed
on behalf of the pagedaemon when it's unlocked. this flag is very similar
to PG_RELEASED, but unlike PG_RELEASED, PG_PAGEOUT can be cleared if the
pageout fails due to eg. an indirect-block buffer being locked.
this allows us to remove the "version" field from struct vm_page,
and together with shrinking "loan_count" from 32 bits to 16,
struct vm_page is now 4 bytes smaller.
- no longer use PG_RELEASED for swap-backed pages. if the page is busy
because it's being paged out, we can't release the swap slot to be
reallocated until that write is complete, but unlike with vnodes we
don't keep a count of in-progress writes so there's no good way to
know when the write is done. instead, when we need to free a busy
swap-backed page, just sleep until we can get it busy ourselves.
- implement a fast-path for extending writes which allows us to avoid
zeroing new pages. this substantially reduces cpu usage.
- encapsulate the data used by the genfs code in a struct genfs_node,
which must be the first element of the filesystem-specific vnode data
for filesystems which use genfs_{get,put}pages().
- eliminate many of the UVM pagerops, since they aren't needed anymore
now that the pager "put" operation is a higher-level operation.
- enhance the genfs code to allow NFS to use the genfs_{get,put}pages
instead of a modified copy.
- clean up struct vnode by removing all the fields that used to be used by
the vfs_cluster.c code (which we don't use anymore with UBC).
- remove kmem_object and mb_object since they were useless.
instead of allocating pages to these objects, we now just allocate
pages with no object. such pages are mapped in the kernel until they
are freed, so we can use the mapping to find the page to free it.
this allows us to remove splvm() protection in several places.
The sum of all these changes improves write throughput on my
decstation 5000/200 to within 1% of the rate of NetBSD 1.5
and reduces the elapsed time for "make release" of a NetBSD 1.5
source tree on my 128MB pc to 10% less than a 1.5 kernel took.
2001-09-16 00:36:31 +04:00
|
|
|
printf("bread(%p, %x, %ld, CRED, %p)=%d\n",
|
|
|
|
vp, lbn, size, bp, error);
|
1998-08-14 07:26:12 +04:00
|
|
|
#endif
|
|
|
|
}
|
2001-02-27 07:37:44 +03:00
|
|
|
n = MIN(n, size - bp->b_resid);
|
1998-08-14 07:26:12 +04:00
|
|
|
if (error) {
|
|
|
|
#ifdef FILECORE_DEBUG_BR
|
1998-08-14 22:04:04 +04:00
|
|
|
printf("brelse(%p) vn1\n", bp);
|
1998-08-14 07:26:12 +04:00
|
|
|
#endif
|
|
|
|
brelse(bp);
|
|
|
|
return (error);
|
|
|
|
}
|
|
|
|
|
|
|
|
error = uiomove(bp->b_data + on, (int)n, uio);
|
|
|
|
#ifdef FILECORE_DEBUG_BR
|
1998-08-14 22:04:04 +04:00
|
|
|
printf("brelse(%p) vn2\n", bp);
|
1998-08-14 07:26:12 +04:00
|
|
|
#endif
|
|
|
|
brelse(bp);
|
|
|
|
} while (error == 0 && uio->uio_resid > 0 && n != 0);
|
2000-11-27 11:39:39 +03:00
|
|
|
|
|
|
|
out:
|
1998-08-14 07:26:12 +04:00
|
|
|
return (error);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Vnode op for readdir
|
|
|
|
*/
|
|
|
|
int
|
|
|
|
filecore_readdir(v)
|
|
|
|
void *v;
|
|
|
|
{
|
|
|
|
struct vop_readdir_args /* {
|
|
|
|
struct vnode *a_vp;
|
|
|
|
struct uio *a_uio;
|
|
|
|
struct ucred *a_cred;
|
|
|
|
int *a_eofflag;
|
|
|
|
off_t **a_cookies;
|
|
|
|
int *a_ncookies;
|
|
|
|
} */ *ap = v;
|
1998-08-14 22:04:04 +04:00
|
|
|
struct uio *uio = ap->a_uio;
|
1998-08-14 07:26:12 +04:00
|
|
|
struct vnode *vdp = ap->a_vp;
|
|
|
|
struct filecore_node *dp;
|
|
|
|
struct filecore_mnt *fcmp;
|
|
|
|
struct buf *bp = NULL;
|
|
|
|
struct dirent de;
|
|
|
|
struct filecore_direntry *dep = NULL;
|
|
|
|
int error = 0;
|
|
|
|
off_t *cookies = NULL;
|
|
|
|
int ncookies = 0;
|
1998-09-18 08:17:44 +04:00
|
|
|
int i;
|
|
|
|
off_t uiooff;
|
1998-08-14 07:26:12 +04:00
|
|
|
|
|
|
|
dp = VTOI(vdp);
|
|
|
|
|
|
|
|
if ((dp->i_dirent.attr & FILECORE_ATTR_DIR) == 0)
|
|
|
|
return (ENOTDIR);
|
|
|
|
|
1998-09-18 08:17:44 +04:00
|
|
|
if (uio->uio_offset % FILECORE_DIRENT_SIZE != 0)
|
1998-08-14 07:26:12 +04:00
|
|
|
return (EINVAL);
|
1998-08-14 22:04:04 +04:00
|
|
|
i = uio->uio_offset / FILECORE_DIRENT_SIZE;
|
1998-09-18 08:17:44 +04:00
|
|
|
uiooff = uio->uio_offset;
|
1998-08-14 07:26:12 +04:00
|
|
|
|
1998-09-18 08:17:44 +04:00
|
|
|
*ap->a_eofflag = 0;
|
1998-08-14 07:26:12 +04:00
|
|
|
fcmp = dp->i_mnt;
|
|
|
|
|
|
|
|
error = filecore_dbread(dp, &bp);
|
|
|
|
if (error) {
|
1999-02-10 16:14:08 +03:00
|
|
|
brelse(bp);
|
1998-08-14 07:26:12 +04:00
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (ap->a_ncookies == NULL)
|
|
|
|
cookies = NULL;
|
|
|
|
else {
|
1998-09-18 08:17:44 +04:00
|
|
|
*ap->a_ncookies = 0;
|
|
|
|
ncookies = uio->uio_resid/16;
|
2000-08-03 07:38:39 +04:00
|
|
|
cookies = malloc(ncookies * sizeof(off_t), M_TEMP, M_WAITOK);
|
1998-08-14 07:26:12 +04:00
|
|
|
}
|
|
|
|
|
1998-09-18 08:17:44 +04:00
|
|
|
for (; ; i++) {
|
1998-08-14 07:26:12 +04:00
|
|
|
switch (i) {
|
|
|
|
case 0:
|
|
|
|
/* Fake the '.' entry */
|
|
|
|
de.d_fileno = dp->i_number;
|
|
|
|
de.d_type = DT_DIR;
|
|
|
|
de.d_namlen = 1;
|
|
|
|
strcpy(de.d_name, ".");
|
|
|
|
break;
|
|
|
|
case 1:
|
|
|
|
/* Fake the '..' entry */
|
|
|
|
de.d_fileno = filecore_getparent(dp);
|
|
|
|
de.d_type = DT_DIR;
|
|
|
|
de.d_namlen = 2;
|
|
|
|
strcpy(de.d_name, "..");
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
de.d_fileno = dp->i_dirent.addr +
|
1998-08-14 22:04:04 +04:00
|
|
|
((i - 2) << FILECORE_INO_INDEX);
|
|
|
|
dep = fcdirentry(bp->b_data, i - 2);
|
1998-08-14 07:26:12 +04:00
|
|
|
if (dep->attr & FILECORE_ATTR_DIR)
|
|
|
|
de.d_type = DT_DIR;
|
|
|
|
else
|
|
|
|
de.d_type = DT_REG;
|
|
|
|
if (filecore_fn2unix(dep->name, de.d_name,
|
1998-09-18 08:17:44 +04:00
|
|
|
&de.d_namlen)) {
|
|
|
|
*ap->a_eofflag = 1;
|
1998-08-14 07:26:12 +04:00
|
|
|
goto out;
|
1998-09-18 08:17:44 +04:00
|
|
|
}
|
1998-08-14 07:26:12 +04:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
de.d_reclen = DIRENT_SIZE(&de);
|
|
|
|
if (uio->uio_resid < de.d_reclen)
|
|
|
|
goto out;
|
|
|
|
error = uiomove((caddr_t) &de, de.d_reclen, uio);
|
|
|
|
if (error)
|
|
|
|
goto out;
|
1998-09-18 08:17:44 +04:00
|
|
|
uiooff += FILECORE_DIRENT_SIZE;
|
1998-08-14 07:26:12 +04:00
|
|
|
|
|
|
|
if (cookies) {
|
|
|
|
*cookies++ = i*FILECORE_DIRENT_SIZE;
|
|
|
|
(*ap->a_ncookies)++;
|
1998-09-18 08:17:44 +04:00
|
|
|
if (--ncookies == 0) goto out;
|
|
|
|
}
|
1998-08-14 07:26:12 +04:00
|
|
|
}
|
|
|
|
out:
|
1998-09-18 08:17:44 +04:00
|
|
|
if (cookies) {
|
|
|
|
*ap->a_cookies = cookies;
|
1998-08-14 07:26:12 +04:00
|
|
|
if (error) {
|
2000-08-03 07:38:39 +04:00
|
|
|
free(cookies, M_TEMP);
|
1998-08-14 07:26:12 +04:00
|
|
|
*ap->a_ncookies = 0;
|
|
|
|
*ap->a_cookies = NULL;
|
|
|
|
}
|
|
|
|
}
|
1998-09-18 08:17:44 +04:00
|
|
|
uio->uio_offset = uiooff;
|
1998-08-14 07:26:12 +04:00
|
|
|
|
|
|
|
#ifdef FILECORE_DEBUG_BR
|
1998-08-14 22:04:04 +04:00
|
|
|
printf("brelse(%p) vn3\n", bp);
|
1998-08-14 07:26:12 +04:00
|
|
|
#endif
|
|
|
|
brelse (bp);
|
|
|
|
|
|
|
|
return (error);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Return target name of a symbolic link
|
|
|
|
* Shouldn't we get the parent vnode and read the data from there?
|
|
|
|
* This could eventually result in deadlocks in filecore_lookup.
|
|
|
|
* But otherwise the block read here is in the block buffer two times.
|
|
|
|
*/
|
|
|
|
int
|
|
|
|
filecore_readlink(v)
|
|
|
|
void *v;
|
|
|
|
{
|
|
|
|
#if 0
|
|
|
|
struct vop_readlink_args /* {
|
|
|
|
struct vnode *a_vp;
|
|
|
|
struct uio *a_uio;
|
|
|
|
struct ucred *a_cred;
|
|
|
|
} */ *ap = v;
|
|
|
|
#endif
|
1998-09-18 08:17:44 +04:00
|
|
|
|
1998-08-14 07:26:12 +04:00
|
|
|
return (EINVAL);
|
|
|
|
}
|
|
|
|
|
|
|
|
int
|
|
|
|
filecore_link(v)
|
|
|
|
void *v;
|
|
|
|
{
|
|
|
|
struct vop_link_args /* {
|
|
|
|
struct vnode *a_dvp;
|
|
|
|
struct vnode *a_vp;
|
|
|
|
struct componentname *a_cnp;
|
|
|
|
} */ *ap = v;
|
|
|
|
|
|
|
|
VOP_ABORTOP(ap->a_dvp, ap->a_cnp);
|
|
|
|
vput(ap->a_dvp);
|
|
|
|
return (EROFS);
|
|
|
|
}
|
|
|
|
|
|
|
|
int
|
|
|
|
filecore_symlink(v)
|
|
|
|
void *v;
|
|
|
|
{
|
|
|
|
struct vop_symlink_args /* {
|
|
|
|
struct vnode *a_dvp;
|
|
|
|
struct vnode **a_vpp;
|
|
|
|
struct componentname *a_cnp;
|
|
|
|
struct vattr *a_vap;
|
|
|
|
char *a_target;
|
|
|
|
} */ *ap = v;
|
|
|
|
|
|
|
|
VOP_ABORTOP(ap->a_dvp, ap->a_cnp);
|
|
|
|
vput(ap->a_dvp);
|
|
|
|
return (EROFS);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Calculate the logical to physical mapping if not done already,
|
|
|
|
* then call the device strategy routine.
|
|
|
|
*/
|
|
|
|
int
|
|
|
|
filecore_strategy(v)
|
|
|
|
void *v;
|
|
|
|
{
|
|
|
|
struct vop_strategy_args /* {
|
|
|
|
struct buf *a_bp;
|
|
|
|
} */ *ap = v;
|
1998-08-14 22:04:04 +04:00
|
|
|
struct buf *bp = ap->a_bp;
|
|
|
|
struct vnode *vp = bp->b_vp;
|
|
|
|
struct filecore_node *ip;
|
1998-08-14 07:26:12 +04:00
|
|
|
int error;
|
|
|
|
|
|
|
|
ip = VTOI(vp);
|
|
|
|
if (bp->b_blkno == bp->b_lblkno) {
|
|
|
|
error = VOP_BMAP(vp, bp->b_lblkno, NULL, &bp->b_blkno, NULL);
|
|
|
|
if (error) {
|
|
|
|
bp->b_error = error;
|
|
|
|
bp->b_flags |= B_ERROR;
|
|
|
|
biodone(bp);
|
|
|
|
return (error);
|
|
|
|
}
|
|
|
|
if ((long)bp->b_blkno == -1)
|
|
|
|
clrbuf(bp);
|
|
|
|
}
|
|
|
|
if ((long)bp->b_blkno == -1) {
|
|
|
|
biodone(bp);
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
vp = ip->i_devvp;
|
|
|
|
bp->b_dev = vp->v_rdev;
|
|
|
|
VOCALL (vp->v_op, VOFFSET(vop_strategy), ap);
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Print out the contents of an inode.
|
|
|
|
*/
|
|
|
|
/*ARGSUSED*/
|
|
|
|
int
|
|
|
|
filecore_print(v)
|
|
|
|
void *v;
|
|
|
|
{
|
|
|
|
|
|
|
|
printf("tag VT_FILECORE, filecore vnode\n");
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Return POSIX pathconf information applicable to filecore filesystems.
|
|
|
|
*/
|
|
|
|
int
|
|
|
|
filecore_pathconf(v)
|
|
|
|
void *v;
|
|
|
|
{
|
|
|
|
struct vop_pathconf_args /* {
|
|
|
|
struct vnode *a_vp;
|
|
|
|
int a_name;
|
|
|
|
register_t *a_retval;
|
|
|
|
} */ *ap = v;
|
|
|
|
switch (ap->a_name) {
|
|
|
|
case _PC_LINK_MAX:
|
|
|
|
*ap->a_retval = 1;
|
|
|
|
return (0);
|
|
|
|
case _PC_NAME_MAX:
|
|
|
|
*ap->a_retval = 10;
|
|
|
|
return (0);
|
|
|
|
case _PC_PATH_MAX:
|
|
|
|
*ap->a_retval = 256;
|
|
|
|
return (0);
|
|
|
|
case _PC_CHOWN_RESTRICTED:
|
|
|
|
*ap->a_retval = 1;
|
|
|
|
return (0);
|
|
|
|
case _PC_NO_TRUNC:
|
|
|
|
*ap->a_retval = 1;
|
|
|
|
return (0);
|
1998-09-07 19:58:37 +04:00
|
|
|
case _PC_SYNC_IO:
|
|
|
|
*ap->a_retval = 1;
|
|
|
|
return (0);
|
1999-03-22 22:21:07 +03:00
|
|
|
case _PC_FILESIZEBITS:
|
|
|
|
*ap->a_retval = 32;
|
|
|
|
return (0);
|
1998-08-14 07:26:12 +04:00
|
|
|
default:
|
|
|
|
return (EINVAL);
|
|
|
|
}
|
|
|
|
/* NOTREACHED */
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Global vfs data structures for isofs
|
|
|
|
*/
|
|
|
|
#define filecore_create genfs_eopnotsupp
|
|
|
|
#define filecore_mknod genfs_eopnotsupp
|
|
|
|
#define filecore_write genfs_eopnotsupp
|
|
|
|
#define filecore_setattr genfs_eopnotsupp
|
|
|
|
#ifdef NFSSERVER
|
|
|
|
int lease_check __P((void *));
|
|
|
|
#define filecore_lease_check lease_check
|
|
|
|
#else
|
|
|
|
#define filecore_lease_check genfs_nullop
|
|
|
|
#endif
|
1999-08-04 00:19:16 +04:00
|
|
|
#define filecore_fcntl genfs_fcntl
|
1998-08-14 07:26:12 +04:00
|
|
|
#define filecore_ioctl genfs_enoioctl
|
|
|
|
#define filecore_fsync genfs_nullop
|
|
|
|
#define filecore_remove genfs_eopnotsupp
|
|
|
|
#define filecore_rename genfs_eopnotsupp
|
|
|
|
#define filecore_mkdir genfs_eopnotsupp
|
|
|
|
#define filecore_rmdir genfs_eopnotsupp
|
|
|
|
#define filecore_advlock genfs_eopnotsupp
|
|
|
|
#define filecore_valloc genfs_eopnotsupp
|
|
|
|
#define filecore_vfree genfs_nullop
|
|
|
|
#define filecore_truncate genfs_eopnotsupp
|
|
|
|
#define filecore_update genfs_nullop
|
|
|
|
#define filecore_bwrite genfs_eopnotsupp
|
|
|
|
#define filecore_revoke genfs_revoke
|
|
|
|
#define filecore_blkatoff genfs_eopnotsupp
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Global vfs data structures for filecore
|
|
|
|
*/
|
|
|
|
int (**filecore_vnodeop_p) __P((void *));
|
2001-01-22 16:32:23 +03:00
|
|
|
const struct vnodeopv_entry_desc filecore_vnodeop_entries[] = {
|
1998-08-14 07:26:12 +04:00
|
|
|
{ &vop_default_desc, vn_default_error },
|
|
|
|
{ &vop_lookup_desc, filecore_lookup }, /* lookup */
|
|
|
|
{ &vop_create_desc, filecore_create }, /* create */
|
|
|
|
{ &vop_mknod_desc, filecore_mknod }, /* mknod */
|
|
|
|
{ &vop_open_desc, filecore_open }, /* open */
|
|
|
|
{ &vop_close_desc, filecore_close }, /* close */
|
|
|
|
{ &vop_access_desc, filecore_access }, /* access */
|
|
|
|
{ &vop_getattr_desc, filecore_getattr }, /* getattr */
|
|
|
|
{ &vop_setattr_desc, filecore_setattr }, /* setattr */
|
|
|
|
{ &vop_read_desc, filecore_read }, /* read */
|
|
|
|
{ &vop_write_desc, filecore_write }, /* write */
|
|
|
|
{ &vop_lease_desc, filecore_lease_check }, /* lease */
|
1999-08-04 00:19:16 +04:00
|
|
|
{ &vop_fcntl_desc, filecore_fcntl }, /* fcntl */
|
1998-08-14 07:26:12 +04:00
|
|
|
{ &vop_ioctl_desc, filecore_ioctl }, /* ioctl */
|
|
|
|
{ &vop_poll_desc, filecore_poll }, /* poll */
|
|
|
|
{ &vop_revoke_desc, filecore_revoke }, /* revoke */
|
|
|
|
{ &vop_mmap_desc, filecore_mmap }, /* mmap */
|
|
|
|
{ &vop_fsync_desc, filecore_fsync }, /* fsync */
|
|
|
|
{ &vop_seek_desc, filecore_seek }, /* seek */
|
|
|
|
{ &vop_remove_desc, filecore_remove }, /* remove */
|
|
|
|
{ &vop_link_desc, filecore_link }, /* link */
|
|
|
|
{ &vop_rename_desc, filecore_rename }, /* rename */
|
|
|
|
{ &vop_mkdir_desc, filecore_mkdir }, /* mkdir */
|
|
|
|
{ &vop_rmdir_desc, filecore_rmdir }, /* rmdir */
|
|
|
|
{ &vop_symlink_desc, filecore_symlink }, /* symlink */
|
|
|
|
{ &vop_readdir_desc, filecore_readdir }, /* readdir */
|
|
|
|
{ &vop_readlink_desc, filecore_readlink }, /* readlink */
|
|
|
|
{ &vop_abortop_desc, filecore_abortop }, /* abortop */
|
|
|
|
{ &vop_inactive_desc, filecore_inactive }, /* inactive */
|
|
|
|
{ &vop_reclaim_desc, filecore_reclaim }, /* reclaim */
|
1999-07-08 05:05:58 +04:00
|
|
|
{ &vop_lock_desc, genfs_lock }, /* lock */
|
|
|
|
{ &vop_unlock_desc, genfs_unlock }, /* unlock */
|
1998-08-14 07:26:12 +04:00
|
|
|
{ &vop_bmap_desc, filecore_bmap }, /* bmap */
|
|
|
|
{ &vop_strategy_desc, filecore_strategy }, /* strategy */
|
|
|
|
{ &vop_print_desc, filecore_print }, /* print */
|
1999-07-08 05:05:58 +04:00
|
|
|
{ &vop_islocked_desc, genfs_islocked }, /* islocked */
|
1998-08-14 07:26:12 +04:00
|
|
|
{ &vop_pathconf_desc, filecore_pathconf }, /* pathconf */
|
|
|
|
{ &vop_advlock_desc, filecore_advlock }, /* advlock */
|
|
|
|
{ &vop_blkatoff_desc, filecore_blkatoff }, /* blkatoff */
|
|
|
|
{ &vop_valloc_desc, filecore_valloc }, /* valloc */
|
|
|
|
{ &vop_vfree_desc, filecore_vfree }, /* vfree */
|
|
|
|
{ &vop_truncate_desc, filecore_truncate }, /* truncate */
|
|
|
|
{ &vop_update_desc, filecore_update }, /* update */
|
|
|
|
{ &vop_bwrite_desc, vn_bwrite }, /* bwrite */
|
2000-11-27 11:39:39 +03:00
|
|
|
{ &vop_getpages_desc, genfs_getpages }, /* getpages */
|
a whole bunch of changes to improve performance and robustness under load:
- remove special treatment of pager_map mappings in pmaps. this is
required now, since I've removed the globals that expose the address range.
pager_map now uses pmap_kenter_pa() instead of pmap_enter(), so there's
no longer any need to special-case it.
- eliminate struct uvm_vnode by moving its fields into struct vnode.
- rewrite the pageout path. the pager is now responsible for handling the
high-level requests instead of only getting control after a bunch of work
has already been done on its behalf. this will allow us to UBCify LFS,
which needs tighter control over its pages than other filesystems do.
writing a page to disk no longer requires making it read-only, which
allows us to write wired pages without causing all kinds of havoc.
- use a new PG_PAGEOUT flag to indicate that a page should be freed
on behalf of the pagedaemon when it's unlocked. this flag is very similar
to PG_RELEASED, but unlike PG_RELEASED, PG_PAGEOUT can be cleared if the
pageout fails due to eg. an indirect-block buffer being locked.
this allows us to remove the "version" field from struct vm_page,
and together with shrinking "loan_count" from 32 bits to 16,
struct vm_page is now 4 bytes smaller.
- no longer use PG_RELEASED for swap-backed pages. if the page is busy
because it's being paged out, we can't release the swap slot to be
reallocated until that write is complete, but unlike with vnodes we
don't keep a count of in-progress writes so there's no good way to
know when the write is done. instead, when we need to free a busy
swap-backed page, just sleep until we can get it busy ourselves.
- implement a fast-path for extending writes which allows us to avoid
zeroing new pages. this substantially reduces cpu usage.
- encapsulate the data used by the genfs code in a struct genfs_node,
which must be the first element of the filesystem-specific vnode data
for filesystems which use genfs_{get,put}pages().
- eliminate many of the UVM pagerops, since they aren't needed anymore
now that the pager "put" operation is a higher-level operation.
- enhance the genfs code to allow NFS to use the genfs_{get,put}pages
instead of a modified copy.
- clean up struct vnode by removing all the fields that used to be used by
the vfs_cluster.c code (which we don't use anymore with UBC).
- remove kmem_object and mb_object since they were useless.
instead of allocating pages to these objects, we now just allocate
pages with no object. such pages are mapped in the kernel until they
are freed, so we can use the mapping to find the page to free it.
this allows us to remove splvm() protection in several places.
The sum of all these changes improves write throughput on my
decstation 5000/200 to within 1% of the rate of NetBSD 1.5
and reduces the elapsed time for "make release" of a NetBSD 1.5
source tree on my 128MB pc to 10% less than a 1.5 kernel took.
2001-09-16 00:36:31 +04:00
|
|
|
{ &vop_putpages_desc, genfs_putpages }, /* putpages */
|
2000-11-27 11:39:39 +03:00
|
|
|
{ NULL, NULL }
|
1998-08-14 07:26:12 +04:00
|
|
|
};
|
2001-01-22 16:32:23 +03:00
|
|
|
const struct vnodeopv_desc filecore_vnodeop_opv_desc =
|
1998-08-14 07:26:12 +04:00
|
|
|
{ &filecore_vnodeop_p, filecore_vnodeop_entries };
|