NetBSD/sys/miscfs/specfs/spec_vnops.c

828 lines
20 KiB
C
Raw Normal View History

/* $NetBSD: spec_vnops.c,v 1.66 2002/10/26 13:50:17 jdolecek Exp $ */
1993-03-21 12:45:37 +03:00
/*
* Copyright (c) 1989, 1993
* The Regents of the University of California. All rights reserved.
1993-03-21 12:45:37 +03:00
*
* 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.
*
1998-03-01 05:20:01 +03:00
* @(#)spec_vnops.c 8.15 (Berkeley) 7/14/95
1993-03-21 12:45:37 +03:00
*/
2001-11-10 16:33:40 +03:00
#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: spec_vnops.c,v 1.66 2002/10/26 13:50:17 jdolecek Exp $");
2001-11-10 16:33:40 +03:00
1993-12-18 06:59:02 +03:00
#include <sys/param.h>
#include <sys/proc.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/conf.h>
#include <sys/buf.h>
#include <sys/mount.h>
#include <sys/namei.h>
#include <sys/vnode.h>
#include <sys/stat.h>
#include <sys/errno.h>
#include <sys/ioctl.h>
#include <sys/file.h>
#include <sys/disklabel.h>
1997-04-02 21:09:47 +04:00
#include <sys/lockf.h>
1996-02-10 01:39:56 +03:00
#include <miscfs/genfs/genfs.h>
#include <miscfs/specfs/specdev.h>
1993-03-21 12:45:37 +03:00
/* symbolic sleep message strings for devices */
const char devopn[] = "devopn";
const char devio[] = "devio";
const char devwait[] = "devwait";
const char devin[] = "devin";
const char devout[] = "devout";
const char devioc[] = "devioc";
const char devcls[] = "devcls";
1993-03-21 12:45:37 +03:00
2002-05-13 00:42:03 +04:00
struct vnode *speclisth[SPECHSZ];
/*
* This vnode operations vector is used for two things only:
* - special device nodes created from whole cloth by the kernel.
* - as a temporary vnodeops replacement for vnodes which were found to
* be aliased by callers of checkalias().
* For the ops vector for vnodes built from special devices found in a
* filesystem, see (e.g) ffs_specop_entries[] in ffs_vnops.c or the
* equivalent for other filesystems.
*/
1996-02-10 01:39:56 +03:00
int (**spec_vnodeop_p) __P((void *));
const struct vnodeopv_entry_desc spec_vnodeop_entries[] = {
{ &vop_default_desc, vn_default_error },
{ &vop_lookup_desc, spec_lookup }, /* lookup */
{ &vop_create_desc, spec_create }, /* create */
{ &vop_mknod_desc, spec_mknod }, /* mknod */
{ &vop_open_desc, spec_open }, /* open */
{ &vop_close_desc, spec_close }, /* close */
{ &vop_access_desc, spec_access }, /* access */
{ &vop_getattr_desc, spec_getattr }, /* getattr */
{ &vop_setattr_desc, spec_setattr }, /* setattr */
{ &vop_read_desc, spec_read }, /* read */
{ &vop_write_desc, spec_write }, /* write */
{ &vop_lease_desc, spec_lease_check }, /* lease */
{ &vop_fcntl_desc, spec_fcntl }, /* fcntl */
{ &vop_ioctl_desc, spec_ioctl }, /* ioctl */
1996-09-07 16:40:22 +04:00
{ &vop_poll_desc, spec_poll }, /* poll */
{ &vop_kqfilter_desc, spec_kqfilter }, /* kqfilter */
1998-03-01 05:20:01 +03:00
{ &vop_revoke_desc, spec_revoke }, /* revoke */
{ &vop_mmap_desc, spec_mmap }, /* mmap */
{ &vop_fsync_desc, spec_fsync }, /* fsync */
{ &vop_seek_desc, spec_seek }, /* seek */
{ &vop_remove_desc, spec_remove }, /* remove */
{ &vop_link_desc, spec_link }, /* link */
{ &vop_rename_desc, spec_rename }, /* rename */
{ &vop_mkdir_desc, spec_mkdir }, /* mkdir */
{ &vop_rmdir_desc, spec_rmdir }, /* rmdir */
{ &vop_symlink_desc, spec_symlink }, /* symlink */
{ &vop_readdir_desc, spec_readdir }, /* readdir */
{ &vop_readlink_desc, spec_readlink }, /* readlink */
{ &vop_abortop_desc, spec_abortop }, /* abortop */
{ &vop_inactive_desc, spec_inactive }, /* inactive */
{ &vop_reclaim_desc, spec_reclaim }, /* reclaim */
{ &vop_lock_desc, spec_lock }, /* lock */
{ &vop_unlock_desc, spec_unlock }, /* unlock */
{ &vop_bmap_desc, spec_bmap }, /* bmap */
{ &vop_strategy_desc, spec_strategy }, /* strategy */
{ &vop_print_desc, spec_print }, /* print */
{ &vop_islocked_desc, spec_islocked }, /* islocked */
{ &vop_pathconf_desc, spec_pathconf }, /* pathconf */
{ &vop_advlock_desc, spec_advlock }, /* advlock */
{ &vop_blkatoff_desc, spec_blkatoff }, /* blkatoff */
{ &vop_valloc_desc, spec_valloc }, /* valloc */
{ &vop_vfree_desc, spec_vfree }, /* vfree */
{ &vop_truncate_desc, spec_truncate }, /* truncate */
{ &vop_update_desc, spec_update }, /* update */
1996-02-10 01:39:56 +03:00
{ &vop_bwrite_desc, spec_bwrite }, /* bwrite */
{ &vop_getpages_desc, spec_getpages }, /* getpages */
{ &vop_putpages_desc, spec_putpages }, /* putpages */
{ NULL, NULL }
1993-03-21 12:45:37 +03:00
};
const struct vnodeopv_desc spec_vnodeop_opv_desc =
{ &spec_vnodeop_p, spec_vnodeop_entries };
1993-03-21 12:45:37 +03:00
/*
* Trivial lookup routine that always fails.
*/
int
1996-02-10 01:39:56 +03:00
spec_lookup(v)
void *v;
{
struct vop_lookup_args /* {
struct vnode *a_dvp;
struct vnode **a_vpp;
struct componentname *a_cnp;
1996-02-10 01:39:56 +03:00
} */ *ap = v;
1993-03-21 12:45:37 +03:00
*ap->a_vpp = NULL;
1993-03-21 12:45:37 +03:00
return (ENOTDIR);
}
/*
* Returns true if dev is /dev/mem or /dev/kmem.
*/
static int
iskmemdev(dev_t dev)
{
/* mem_no is emitted by config(8) to generated devsw.c */
extern const int mem_no;
/* minor 14 is /dev/io on i386 with COMPAT_10 */
return (major(dev) == mem_no && (minor(dev) < 2 || minor(dev) == 14));
}
1993-03-21 12:45:37 +03:00
/*
* Open a special file.
1993-03-21 12:45:37 +03:00
*/
/* ARGSUSED */
1996-02-10 01:39:56 +03:00
int
spec_open(v)
void *v;
{
struct vop_open_args /* {
struct vnode *a_vp;
int a_mode;
struct ucred *a_cred;
struct proc *a_p;
1996-02-10 01:39:56 +03:00
} */ *ap = v;
1998-03-01 05:20:01 +03:00
struct proc *p = ap->a_p;
struct vnode *bvp, *vp = ap->a_vp;
const struct bdevsw *bdev;
const struct cdevsw *cdev;
dev_t blkdev, dev = (dev_t)vp->v_rdev;
1993-03-21 12:45:37 +03:00
int error;
struct partinfo pi;
1993-03-21 12:45:37 +03:00
/*
* Don't allow open if fs is mounted -nodev.
*/
1993-03-21 12:45:37 +03:00
if (vp->v_mount && (vp->v_mount->mnt_flag & MNT_NODEV))
return (ENXIO);
switch (vp->v_type) {
case VCHR:
cdev = cdevsw_lookup(dev);
if (cdev == NULL)
1993-03-21 12:45:37 +03:00
return (ENXIO);
if (ap->a_cred != FSCRED && (ap->a_mode & FWRITE)) {
/*
* When running in very secure mode, do not allow
* opens for writing of any disk character devices.
*/
if (securelevel >= 2 && cdev->d_type == D_DISK)
return (EPERM);
/*
* When running in secure mode, do not allow opens
* for writing of /dev/mem, /dev/kmem, or character
* devices whose corresponding block devices are
* currently mounted.
*/
if (securelevel >= 1) {
blkdev = devsw_chr2blk(dev);
if (blkdev != (dev_t)NODEV &&
vfinddev(blkdev, VBLK, &bvp) &&
(error = vfs_mountedon(bvp)))
return (error);
if (iskmemdev(dev))
return (EPERM);
}
}
if (cdev->d_type == D_TTY)
vp->v_flag |= VISTTY;
1998-03-01 05:20:01 +03:00
VOP_UNLOCK(vp, 0);
error = (*cdev->d_open)(dev, ap->a_mode, S_IFCHR, p);
1998-03-01 05:20:01 +03:00
vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
1993-03-21 12:45:37 +03:00
return (error);
case VBLK:
bdev = bdevsw_lookup(dev);
if (bdev == NULL)
1993-03-21 12:45:37 +03:00
return (ENXIO);
/*
* When running in very secure mode, do not allow
* opens for writing of any disk block devices.
*/
if (securelevel >= 2 && ap->a_cred != FSCRED &&
(ap->a_mode & FWRITE) && bdev->d_type == D_DISK)
return (EPERM);
/*
* Do not allow opens of block devices that are
* currently mounted.
*/
1996-02-10 01:39:56 +03:00
if ((error = vfs_mountedon(vp)) != 0)
1993-03-21 12:45:37 +03:00
return (error);
error = (*bdev->d_open)(dev, ap->a_mode, S_IFBLK, p);
if (error) {
return error;
}
error = (*bdev->d_ioctl)(vp->v_rdev,
DIOCGPART, (caddr_t)&pi, FREAD, curproc);
if (error == 0) {
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
vp->v_size = (voff_t)pi.disklab->d_secsize *
pi.part->p_size;
}
return 0;
1996-02-10 01:39:56 +03:00
case VNON:
case VLNK:
case VDIR:
case VREG:
case VBAD:
case VFIFO:
case VSOCK:
break;
1993-03-21 12:45:37 +03:00
}
return (0);
1993-03-21 12:45:37 +03:00
}
/*
* Vnode op for read
*/
/* ARGSUSED */
1996-02-10 01:39:56 +03:00
int
spec_read(v)
void *v;
{
struct vop_read_args /* {
struct vnode *a_vp;
struct uio *a_uio;
int a_ioflag;
struct ucred *a_cred;
1996-02-10 01:39:56 +03:00
} */ *ap = v;
2000-03-30 16:22:12 +04:00
struct vnode *vp = ap->a_vp;
struct uio *uio = ap->a_uio;
struct proc *p = uio->uio_procp;
struct buf *bp;
const struct bdevsw *bdev;
const struct cdevsw *cdev;
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
daddr_t bn;
int bsize, bscale;
struct partinfo dpart;
int n, on;
1993-03-21 12:45:37 +03:00
int error = 0;
#ifdef DIAGNOSTIC
if (uio->uio_rw != UIO_READ)
panic("spec_read mode");
if (uio->uio_segflg == UIO_USERSPACE && uio->uio_procp != curproc)
panic("spec_read proc");
#endif
if (uio->uio_resid == 0)
return (0);
switch (vp->v_type) {
case VCHR:
1998-03-01 05:20:01 +03:00
VOP_UNLOCK(vp, 0);
cdev = cdevsw_lookup(vp->v_rdev);
if (cdev != NULL)
error = (*cdev->d_read)(vp->v_rdev, uio, ap->a_ioflag);
else
error = ENXIO;
vn_lock(vp, LK_SHARED | LK_RETRY);
1993-03-21 12:45:37 +03:00
return (error);
case VBLK:
if (uio->uio_offset < 0)
return (EINVAL);
bsize = BLKDEV_IOSIZE;
bdev = bdevsw_lookup(vp->v_rdev);
if (bdev != NULL &&
(*bdev->d_ioctl)(vp->v_rdev, DIOCGPART, (caddr_t)&dpart,
FREAD, p) == 0) {
if (dpart.part->p_fstype == FS_BSDFFS &&
dpart.part->p_frag != 0 && dpart.part->p_fsize != 0)
bsize = dpart.part->p_frag *
dpart.part->p_fsize;
}
bscale = bsize >> DEV_BSHIFT;
do {
bn = (uio->uio_offset >> DEV_BSHIFT) &~ (bscale - 1);
on = uio->uio_offset % bsize;
n = min((unsigned)(bsize - on), uio->uio_resid);
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, bn, bsize, NOCRED, &bp);
n = min(n, bsize - bp->b_resid);
if (error) {
brelse(bp);
return (error);
}
error = uiomove((char *)bp->b_data + on, n, uio);
brelse(bp);
} while (error == 0 && uio->uio_resid > 0 && n != 0);
return (error);
default:
panic("spec_read type");
1993-03-21 12:45:37 +03:00
}
/* NOTREACHED */
1993-03-21 12:45:37 +03:00
}
/*
* Vnode op for write
*/
/* ARGSUSED */
1996-02-10 01:39:56 +03:00
int
spec_write(v)
void *v;
{
struct vop_write_args /* {
struct vnode *a_vp;
struct uio *a_uio;
int a_ioflag;
struct ucred *a_cred;
1996-02-10 01:39:56 +03:00
} */ *ap = v;
2000-03-30 16:22:12 +04:00
struct vnode *vp = ap->a_vp;
struct uio *uio = ap->a_uio;
struct proc *p = uio->uio_procp;
struct buf *bp;
const struct bdevsw *bdev;
const struct cdevsw *cdev;
daddr_t bn;
int bsize, bscale;
struct partinfo dpart;
int n, on;
1993-03-21 12:45:37 +03:00
int error = 0;
#ifdef DIAGNOSTIC
if (uio->uio_rw != UIO_WRITE)
panic("spec_write mode");
if (uio->uio_segflg == UIO_USERSPACE && uio->uio_procp != curproc)
panic("spec_write proc");
#endif
switch (vp->v_type) {
case VCHR:
1998-03-01 05:20:01 +03:00
VOP_UNLOCK(vp, 0);
cdev = cdevsw_lookup(vp->v_rdev);
if (cdev != NULL)
error = (*cdev->d_write)(vp->v_rdev, uio, ap->a_ioflag);
else
error = ENXIO;
1998-03-01 05:20:01 +03:00
vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
1993-03-21 12:45:37 +03:00
return (error);
case VBLK:
if (uio->uio_resid == 0)
return (0);
if (uio->uio_offset < 0)
return (EINVAL);
bsize = BLKDEV_IOSIZE;
bdev = bdevsw_lookup(vp->v_rdev);
if (bdev != NULL &&
(*bdev->d_ioctl)(vp->v_rdev, DIOCGPART, (caddr_t)&dpart,
FREAD, p) == 0) {
if (dpart.part->p_fstype == FS_BSDFFS &&
dpart.part->p_frag != 0 && dpart.part->p_fsize != 0)
bsize = dpart.part->p_frag *
dpart.part->p_fsize;
}
bscale = bsize >> DEV_BSHIFT;
do {
bn = (uio->uio_offset >> DEV_BSHIFT) &~ (bscale - 1);
on = uio->uio_offset % bsize;
n = min((unsigned)(bsize - on), uio->uio_resid);
if (n == bsize)
bp = getblk(vp, bn, bsize, 0, 0);
else
error = bread(vp, bn, bsize, NOCRED, &bp);
if (error) {
brelse(bp);
return (error);
}
n = min(n, bsize - bp->b_resid);
error = uiomove((char *)bp->b_data + on, n, uio);
if (error)
brelse(bp);
else {
if (n + on == bsize)
bawrite(bp);
else
bdwrite(bp);
if (bp->b_flags & B_ERROR)
error = bp->b_error;
}
} while (error == 0 && uio->uio_resid > 0 && n != 0);
return (error);
default:
panic("spec_write type");
}
/* NOTREACHED */
1993-03-21 12:45:37 +03:00
}
/*
* Device ioctl operation.
*/
/* ARGSUSED */
1996-02-10 01:39:56 +03:00
int
spec_ioctl(v)
void *v;
{
struct vop_ioctl_args /* {
struct vnode *a_vp;
u_long a_command;
caddr_t a_data;
int a_fflag;
struct ucred *a_cred;
struct proc *a_p;
1996-02-10 01:39:56 +03:00
} */ *ap = v;
const struct bdevsw *bdev;
const struct cdevsw *cdev;
dev_t dev = ap->a_vp->v_rdev;
1993-03-21 12:45:37 +03:00
switch (ap->a_vp->v_type) {
1993-03-21 12:45:37 +03:00
case VCHR:
cdev = cdevsw_lookup(dev);
if (cdev == NULL)
return (ENXIO);
return ((*cdev->d_ioctl)(dev, ap->a_command, ap->a_data,
ap->a_fflag, ap->a_p));
1993-03-21 12:45:37 +03:00
case VBLK:
bdev = bdevsw_lookup(dev);
if (bdev == NULL)
return (ENXIO);
if (ap->a_command == 0 && (long)ap->a_data == B_TAPE) {
if (bdev->d_type == D_TAPE)
1993-03-21 12:45:37 +03:00
return (0);
else
return (1);
}
return ((*bdev->d_ioctl)(dev, ap->a_command, ap->a_data,
ap->a_fflag, ap->a_p));
1993-03-21 12:45:37 +03:00
default:
panic("spec_ioctl");
/* NOTREACHED */
}
}
/* ARGSUSED */
1996-02-10 01:39:56 +03:00
int
1996-09-07 16:40:22 +04:00
spec_poll(v)
1996-02-10 01:39:56 +03:00
void *v;
{
1996-09-07 16:40:22 +04:00
struct vop_poll_args /* {
struct vnode *a_vp;
1996-09-07 16:40:22 +04:00
int a_events;
struct proc *a_p;
1996-02-10 01:39:56 +03:00
} */ *ap = v;
const struct cdevsw *cdev;
2000-03-30 16:22:12 +04:00
dev_t dev;
1993-03-21 12:45:37 +03:00
switch (ap->a_vp->v_type) {
1993-03-21 12:45:37 +03:00
case VCHR:
dev = ap->a_vp->v_rdev;
cdev = cdevsw_lookup(dev);
if (cdev == NULL)
return (ENXIO);
return (*cdev->d_poll)(dev, ap->a_events, ap->a_p);
default:
1996-09-07 16:40:22 +04:00
return (genfs_poll(v));
1993-03-21 12:45:37 +03:00
}
}
/* ARGSUSED */
int
spec_kqfilter(v)
void *v;
{
struct vop_kqfilter_args /* {
struct vnode *a_vp;
struct proc *a_kn;
} */ *ap = v;
const struct cdevsw *cdev;
dev_t dev;
switch (ap->a_vp->v_type) {
case VCHR:
dev = ap->a_vp->v_rdev;
cdev = cdevsw_lookup(dev);
if (cdev == NULL)
return (ENXIO);
return (*cdev->d_kqfilter)(dev, ap->a_kn);
default:
/*
* Block devices don't support kqfilter, and refuse it
* for any other files (like those vflush()ed) too.
*/
return (EOPNOTSUPP);
}
}
1993-03-21 12:45:37 +03:00
/*
* Synch buffers associated with a block device
1993-03-21 12:45:37 +03:00
*/
/* ARGSUSED */
int
1996-02-10 01:39:56 +03:00
spec_fsync(v)
void *v;
{
struct vop_fsync_args /* {
struct vnode *a_vp;
struct ucred *a_cred;
int a_flags;
2000-09-20 02:01:59 +04:00
off_t offlo;
off_t offhi;
struct proc *a_p;
1996-02-10 01:39:56 +03:00
} */ *ap = v;
2000-03-30 16:22:12 +04:00
struct vnode *vp = ap->a_vp;
if (vp->v_type == VBLK)
vflushbuf(vp, (ap->a_flags & FSYNC_WAIT) != 0);
return (0);
}
/*
* Just call the device strategy routine
*/
1996-02-10 01:39:56 +03:00
int
spec_strategy(v)
void *v;
{
struct vop_strategy_args /* {
struct buf *a_bp;
1996-02-10 01:39:56 +03:00
} */ *ap = v;
struct buf *bp;
const struct bdevsw *bdev;
1993-03-21 12:45:37 +03:00
bp = ap->a_bp;
if (!(bp->b_flags & B_READ) &&
(LIST_FIRST(&bp->b_dep)) != NULL && bioops.io_start)
(*bioops.io_start)(bp);
bdev = bdevsw_lookup(bp->b_dev);
if (bdev != NULL)
(*bdev->d_strategy)(bp);
1993-03-21 12:45:37 +03:00
return (0);
}
1998-03-01 05:20:01 +03:00
int
spec_inactive(v)
void *v;
{
struct vop_inactive_args /* {
struct vnode *a_vp;
struct proc *a_p;
} */ *ap = v;
VOP_UNLOCK(ap->a_vp, 0);
return (0);
}
1993-03-21 12:45:37 +03:00
/*
* This is a noop, simply returning what one has been given.
*/
1996-02-10 01:39:56 +03:00
int
spec_bmap(v)
void *v;
{
struct vop_bmap_args /* {
struct vnode *a_vp;
daddr_t a_bn;
struct vnode **a_vpp;
daddr_t *a_bnp;
1998-03-01 05:20:01 +03:00
int *a_runp;
1996-02-10 01:39:56 +03:00
} */ *ap = v;
1993-03-21 12:45:37 +03:00
if (ap->a_vpp != NULL)
*ap->a_vpp = ap->a_vp;
if (ap->a_bnp != NULL)
*ap->a_bnp = ap->a_bn;
1998-03-01 05:20:01 +03:00
if (ap->a_runp != NULL)
*ap->a_runp = (MAXBSIZE >> DEV_BSHIFT) - 1;
1993-03-21 12:45:37 +03:00
return (0);
}
/*
* Device close routine
*/
/* ARGSUSED */
1996-02-10 01:39:56 +03:00
int
spec_close(v)
void *v;
{
struct vop_close_args /* {
struct vnode *a_vp;
int a_fflag;
struct ucred *a_cred;
struct proc *a_p;
1996-02-10 01:39:56 +03:00
} */ *ap = v;
2000-03-30 16:22:12 +04:00
struct vnode *vp = ap->a_vp;
const struct bdevsw *bdev;
const struct cdevsw *cdev;
1993-03-21 12:45:37 +03:00
dev_t dev = vp->v_rdev;
int (*devclose) __P((dev_t, int, int, struct proc *));
int mode, error, count, flags, flags1;
count = vcount(vp);
simple_lock(&vp->v_interlock);
flags = vp->v_flag;
simple_unlock(&vp->v_interlock);
1993-03-21 12:45:37 +03:00
switch (vp->v_type) {
case VCHR:
/*
* Hack: a tty device that is a controlling terminal
* has a reference from the session structure.
* We cannot easily tell that a character device is
* a controlling terminal, unless it is the closing
* process' controlling terminal. In that case,
* if the reference count is 2 (this last descriptor
* plus the session), release the reference from the session.
*/
if (count == 2 && ap->a_p &&
vp == ap->a_p->p_session->s_ttyvp) {
vrele(vp);
count--;
ap->a_p->p_session->s_ttyvp = NULL;
}
1993-03-21 12:45:37 +03:00
/*
* If the vnode is locked, then we are in the midst
* of forcably closing the device, otherwise we only
* close on last reference.
*/
if (count > 1 && (flags & VXLOCK) == 0)
1993-03-21 12:45:37 +03:00
return (0);
cdev = cdevsw_lookup(dev);
if (cdev != NULL)
devclose = cdev->d_close;
else
devclose = NULL;
1993-03-21 12:45:37 +03:00
mode = S_IFCHR;
break;
case VBLK:
/*
* On last close of a block device (that isn't mounted)
* we must invalidate any in core blocks, so that
* we can, for instance, change floppy disks.
*/
1996-02-10 01:39:56 +03:00
error = vinvalbuf(vp, V_SAVE, ap->a_cred, ap->a_p, 0, 0);
if (error)
return (error);
1993-03-21 12:45:37 +03:00
/*
* We do not want to really close the device if it
* is still in use unless we are trying to close it
* forcibly. Since every use (buffer, vnode, swap, cmap)
* holds a reference to the vnode, and because we mark
* any other vnodes that alias this device, when the
* sum of the reference counts on all the aliased
* vnodes descends to one, we are on last close.
*/
if (count > 1 && (flags & VXLOCK) == 0)
1993-03-21 12:45:37 +03:00
return (0);
bdev = bdevsw_lookup(dev);
if (bdev != NULL)
devclose = bdev->d_close;
else
devclose = NULL;
1993-03-21 12:45:37 +03:00
mode = S_IFBLK;
break;
default:
panic("spec_close: not special");
}
flags1 = ap->a_fflag;
/*
* if VXLOCK is set, then we're going away soon, so make this
* non-blocking. Also ensures that we won't wedge in vn_lock below.
*/
if (flags & VXLOCK)
flags1 |= FNONBLOCK;
/*
2002-07-11 03:16:32 +04:00
* If we're able to block, release the vnode lock & reacquire. We
* might end up sleaping for someone else who wants our queues. They
* won't get them if we hold the vnode locked. Also, if VXLOCK is set,
* don't release the lock as we won't be able to regain it.
*/
if (!(flags1 & FNONBLOCK))
VOP_UNLOCK(vp, 0);
if (devclose != NULL)
error = (*devclose)(dev, flags1, mode, ap->a_p);
else
error = ENXIO;
if (!(flags1 & FNONBLOCK))
vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
return (error);
1993-03-21 12:45:37 +03:00
}
/*
* Print out the contents of a special device vnode.
*/
1996-02-10 01:39:56 +03:00
int
spec_print(v)
void *v;
{
struct vop_print_args /* {
struct vnode *a_vp;
1996-02-10 01:39:56 +03:00
} */ *ap = v;
1993-03-21 12:45:37 +03:00
1996-10-13 06:21:25 +04:00
printf("tag VT_NON, dev %d, %d\n", major(ap->a_vp->v_rdev),
1996-10-11 02:46:11 +04:00
minor(ap->a_vp->v_rdev));
1996-02-10 01:39:56 +03:00
return 0;
}
/*
* Return POSIX pathconf information applicable to special devices.
*/
1996-02-10 01:39:56 +03:00
int
spec_pathconf(v)
void *v;
{
struct vop_pathconf_args /* {
struct vnode *a_vp;
int a_name;
register_t *a_retval;
1996-02-10 01:39:56 +03:00
} */ *ap = v;
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);
1998-08-03 18:19:57 +04:00
case _PC_SYNC_IO:
*ap->a_retval = 1;
return (0);
default:
return (EINVAL);
}
/* NOTREACHED */
1993-03-21 12:45:37 +03:00
}
1997-04-02 21:09:47 +04:00
/*
* Advisory record locking support.
*/
int
spec_advlock(v)
void *v;
{
struct vop_advlock_args /* {
struct vnode *a_vp;
caddr_t a_id;
int a_op;
struct flock *a_fl;
int a_flags;
} */ *ap = v;
2000-03-30 16:22:12 +04:00
struct vnode *vp = ap->a_vp;
1997-04-02 21:09:47 +04:00
return lf_advlock(ap, &vp->v_speclockf, (off_t)0);
1997-04-02 21:09:47 +04:00
}