NetBSD/sys/miscfs/procfs/procfs_subr.c

381 lines
8.8 KiB
C
Raw Normal View History

1993-08-24 16:59:09 +04:00
/*
1993-08-24 20:25:07 +04:00
* Copyright (c) 1993 Paul Kranenburg
* 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 Christopher G. Demetriou.
* 4. The name of the author may not be used to endorse or promote products
* derived from this software withough specific prior written permission
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``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 AUTHOR 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.
*
* $Id: procfs_subr.c,v 1.2 1993/08/24 16:25:10 pk Exp $
1993-08-24 16:59:09 +04:00
*/
#include "param.h"
#include "systm.h"
#include "time.h"
#include "kernel.h"
#include "ioctl.h"
#include "proc.h"
#include "buf.h"
#include "vnode.h"
#include "file.h"
#include "resourcevar.h"
#include "vm/vm.h"
#include "vm/vm_page.h"
#include "vm/vm_kern.h"
#include "kinfo.h"
#include "kinfo_proc.h"
#include "procfs.h"
#include "pfsnode.h"
#include "machine/vmparam.h"
/*
* Get process address map (PIOCGMAP)
*/
int
pfs_vmmap(procp, pfsp, pmapp)
struct proc *procp;
struct nfsnode *pfsp;
struct procmap *pmapp;
{
int error = 0;
vm_map_t map;
vm_map_entry_t entry;
struct procmap prmap;
map = &procp->p_vmspace->vm_map;
vm_map_lock(map);
entry = map->header.next;
while (entry != &map->header) {
if (entry->is_a_map) {
vm_map_t submap = entry->object.share_map;
vm_map_entry_t subentry;
vm_map_lock(submap);
subentry = submap->header.next;
while (subentry != &submap->header) {
prmap.vaddr = subentry->start;
prmap.size = subentry->end - subentry->start;
prmap.offset = subentry->offset;
prmap.prot = subentry->protection;
error = copyout(&prmap, pmapp, sizeof(prmap));
if (error)
break;
pmapp++;
subentry = subentry->next;
}
vm_map_unlock(submap);
if (error)
break;
}
prmap.vaddr = entry->start;
prmap.size = entry->end - entry->start;
prmap.offset = entry->offset;
prmap.prot = entry->protection;
error = copyout(&prmap, pmapp, sizeof(prmap));
if (error)
break;
pmapp++;
entry = entry->next;
}
vm_map_unlock(map);
return error;
}
/*
* Count number of VM entries of process (PIOCNMAP)
*/
int
pfs_vm_nentries(procp, pfsp)
struct proc *procp;
struct nfsnode *pfsp;
{
int count = 0;
vm_map_t map;
vm_map_entry_t entry;
map = &procp->p_vmspace->vm_map;
vm_map_lock(map);
entry = map->header.next;
while (entry != &map->header) {
if (entry->is_a_map)
count += entry->object.share_map->nentries;
else
count++;
entry = entry->next;
}
vm_map_unlock(map);
return count;
}
/*
* Map process mapped file to file descriptor (PIOCGMAPFD)
*/
int
pfs_vmfd(procp, pfsp, vmfdp, p)
struct proc *procp;
struct pfsnode *pfsp;
struct vmfd *vmfdp;
struct proc *p;
{
int rv;
vm_map_t map;
vm_offset_t addr;
vm_size_t size;
vm_prot_t prot, maxprot;
vm_inherit_t inherit;
boolean_t shared;
vm_object_t object;
vm_offset_t objoff;
struct vnode *vp;
struct file *fp;
extern struct fileops vnops;
map = &procp->p_vmspace->vm_map;
addr = vmfdp->vaddr;
rv = vm_region(map, &addr, &size, &prot, &maxprot,
&inherit, &shared, &object, &objoff);
if (rv != KERN_SUCCESS)
return EINVAL;
while (object != NULL && object->pager == NULL)
object = object->shadow;
if (object == NULL || object->pager == NULL
/* Nobody seems to care || !object->pager_ready */ )
return ENOENT;
if (object->pager->pg_type != PG_VNODE)
return ENOENT;
/* We have a vnode pager, allocate file descriptor */
vp = (struct vnode *)object->pager->pg_handle;
if (VOP_ACCESS(vp, VREAD, p->p_ucred, p)) {
rv = EACCES;
goto out;
}
rv = falloc(p, &fp, &vmfdp->fd);
if (rv)
goto out;
VREF(vp);
fp->f_type = DTYPE_VNODE;
fp->f_ops = &vnops;
fp->f_data = (caddr_t)vp;
fp->f_flag = FREAD;
out:
vm_object_unlock(object);
return rv;
}
/*
* Vnode op for reading/writing.
*/
/* ARGSUSED */
pfs_doio(vp, uio, ioflag, cred)
struct vnode *vp;
register struct uio *uio;
int ioflag;
struct ucred *cred;
{
struct pfsnode *pfsp = VTOPFS(vp);
struct proc *procp;
int error = 0;
long n, on;
#ifdef DEBUG
if (pfs_debug)
printf("pfs_doio(%s): vp 0x%x, proc %x\n",
uio->uio_rw==UIO_READ?"R":"W", vp, uio->uio_procp);
#endif
#ifdef DIAGNOSTIC
if (vp->v_type != VPROC)
panic("pfs_doio vtype");
#endif
procp = pfsp->pfs_pid?pfind(pfsp->pfs_pid):&proc0;
if (!procp)
return ESRCH;
if (procp->p_flag & SSYS)
return EACCES;
if (uio->uio_resid == 0)
return (0);
if (uio->uio_offset < 0)
return (EINVAL);
do { /* One page at a time */
int rv;
vm_map_t map;
vm_offset_t offset;
vm_size_t size;
vm_prot_t oldprot = 0, prot, maxprot;
vm_inherit_t inherit;
boolean_t shared;
vm_object_t object;
vm_offset_t objoff;
vm_page_t m;
vm_offset_t kva;
on = uio->uio_offset - trunc_page(uio->uio_offset);
n = MIN(PAGE_SIZE-on, uio->uio_resid);
/* Map page into kernel space */
map = &procp->p_vmspace->vm_map;
#if 0
vm_map_print(map, 1);
#endif
offset = trunc_page(uio->uio_offset);
rv = vm_region(map, &offset, &size, &prot, &maxprot,
&inherit, &shared, &object, &objoff);
if (rv != KERN_SUCCESS)
return EINVAL;
vm_object_unlock(object);
if (uio->uio_rw == UIO_WRITE && (prot & VM_PROT_WRITE) == 0) {
oldprot = prot;
prot |= VM_PROT_WRITE;
rv = vm_protect(map, offset, PAGE_SIZE, FALSE, prot);
if (rv != KERN_SUCCESS)
return EPERM;
}
/* Just fault the page */
rv = vm_fault(map, offset, prot, FALSE);
if (rv != KERN_SUCCESS)
return EFAULT;
/* Look up again as vm_fault() may have inserted a shadow object */
rv = vm_region(map, &offset, &size, &prot, &maxprot,
&inherit, &shared, &object, &objoff);
if (rv != KERN_SUCCESS)
return EINVAL;
/* Now find the page */
/* XXX hope it's still there, should we have wired it? */
m = vm_page_lookup(object, objoff);
if (m == NULL)
return ESRCH;
kva = kmem_alloc_wait(kernel_map, PAGE_SIZE);
pmap_enter(vm_map_pmap(kernel_map), kva, VM_PAGE_TO_PHYS(m),
VM_PROT_DEFAULT, TRUE);
error = uiomove(kva + on, (int)n, uio);
pmap_remove(vm_map_pmap(kernel_map), kva, kva + PAGE_SIZE);
kmem_free_wakeup(kernel_map, kva, PAGE_SIZE);
if (oldprot) {
rv = vm_protect(map, offset, PAGE_SIZE, FALSE, oldprot);
if (rv != KERN_SUCCESS)
return EPERM;
}
} while (error == 0 && uio->uio_resid > 0);
return (error);
}
#if 00
int
pfs_map(procp, kva, rw, offset)
struct proc *procp;
int rw;
vm_offset_t *kva, offset;
{
int rv;
vm_map_t map;
vm_size_t size;
vm_prot_t prot, maxprot;
vm_inherit_t inherit;
boolean_t shared;
vm_object_t object;
vm_offset_t objoff;
vm_page_t m;
map = &procp->p_vmspace->vm_map;
#if 0
vm_map_print(map, 1);
#endif
offset = trunc_page(offset);
rv = vm_region(map, &offset, &size, &prot, &maxprot,
&inherit, &shared, &object, &objoff);
if (rv != KERN_SUCCESS)
return EINVAL;
vm_object_unlock(object);
if (rw == UIO_WRITE && (prot & VM_PROT_WRITE) == 0) {
prot |= VM_PROT_WRITE;
rv = vm_protect(map, offset, PAGE_SIZE, FALSE, prot);
if (rv != KERN_SUCCESS)
return EPERM;
}
/* Just fault page */
rv = vm_fault(map, offset, prot, FALSE);
if (rv != KERN_SUCCESS)
return EFAULT;
/* Look up again as vm_fault() may have inserted a shadow object */
rv = vm_region(map, &offset, &size, &prot, &maxprot,
&inherit, &shared, &object, &objoff);
if (rv != KERN_SUCCESS)
return EINVAL;
m = vm_page_lookup(object, objoff);
if (m == NULL)
return ESRCH;
*kva = kmem_alloc_wait(kernel_map, PAGE_SIZE);
pmap_enter(vm_map_pmap(kernel_map), *kva, VM_PAGE_TO_PHYS(m),
VM_PROT_DEFAULT, TRUE);
return 0;
}
int
pfs_unmap(procp, kva)
struct proc *procp;
vm_offset_t kva;
{
pmap_remove(vm_map_pmap(kernel_map), kva, kva + PAGE_SIZE);
kmem_free_wakeup(kernel_map, kva, PAGE_SIZE);
}
#endif