freebsd-nq/sys/miscfs/devfs/devfs_tree.c
1998-06-21 14:53:44 +00:00

1292 lines
33 KiB
C
Raw Blame History

This file contains invisible Unicode characters

This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

/*
* Copyright 1997,1998 Julian Elischer. All rights reserved.
* julian@freebsd.org
*
* 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.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDER ``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 HOLDER 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.
*
* $Id: devfs_tree.c,v 1.53 1998/04/20 03:57:35 julian Exp $
*/
/* SPLIT_DEVS means each devfs uses a different vnode for the same device */
/* Otherwise the same device always ends up at the same vnode even if */
/* reached througgh a different devfs instance. The practical difference */
/* is that with the same vnode, chmods and chowns show up on all instances of */
/* a device. (etc) */
#define SPLIT_DEVS 1 /* maybe make this an option */
/*#define SPLIT_DEVS 1*/
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/conf.h>
#include <sys/malloc.h>
#include <sys/mount.h>
#include <sys/proc.h>
#include <sys/vnode.h>
#include <sys/devfsext.h>
#include <machine/stdarg.h>
#include <miscfs/devfs/devfsdefs.h>
static MALLOC_DEFINE(M_DEVFSNODE, "DEVFS node", "DEVFS node");
static MALLOC_DEFINE(M_DEVFSNAME, "DEVFS name", "DEVFS name");
devnm_p dev_root; /* root of the backing tree */
struct mount *devfs_hidden_mount;
int devfs_up_and_going;
/*
* Set up the root directory node in the backing plane
* This is happenning before the vfs system has been
* set up yet, so be careful about what we reference..
* Notice that the ops are by indirection.. as they haven't
* been set up yet!
* DEVFS has a hidden mountpoint that is used as the anchor point
* for the internal 'blueprint' version of the dev filesystem tree.
*/
/*proto*/
void
devfs_sinit(void *junk)
{
int retval; /* we will discard this */
/*
* call the right routine at the right time with the right args....
*/
retval = dev_add_entry("root", NULL, DEV_DIR, NULL, NULL, NULL, &dev_root);
#ifdef PARANOID
if(retval) panic("devfs_sinit: dev_add_entry failed ");
#endif
devfs_hidden_mount = (struct mount *)malloc(sizeof(struct mount),
M_MOUNT,M_NOWAIT);
#ifdef PARANOID
if(!devfs_hidden_mount) panic("devfs_sinit: malloc failed");
#endif
bzero(devfs_hidden_mount,sizeof(struct mount));
devfs_mount(devfs_hidden_mount,"dummy",NULL,NULL,NULL);
dev_root->dnp->dvm = (struct devfsmount *)devfs_hidden_mount->mnt_data;
devfs_up_and_going = 1;
printf("DEVFS: ready for devices\n");
/* part 2 of this is done later */
}
SYSINIT(devfs, SI_SUB_DEVFS, SI_ORDER_FIRST, devfs_sinit, NULL)
/***********************************************************************\
*************************************************************************
* Routines used to find our way to a point in the tree *
*************************************************************************
\***********************************************************************/
/***************************************************************\
* Search down the linked list off a dir to find "name" *
* return the dn_p for that node.
\***************************************************************/
/*proto*/
devnm_p
dev_findname(dn_p dir,char *name)
{
devnm_p newfp;
DBPRINT((" dev_findname(%s)\n",name));
if(dir->type != DEV_DIR) return 0;/*XXX*/ /* printf?*/
if(name[0] == '.')
{
if(name[1] == 0)
{
return dir->by.Dir.myname;
}
if((name[1] == '.') && (name[2] == 0))
{
/* for root, .. == . */
return dir->by.Dir.parent->by.Dir.myname;
}
}
newfp = dir->by.Dir.dirlist;
while(newfp)
{
if(!(strcmp(name,newfp->name)))
return newfp;
newfp = newfp->next;
}
return NULL;
}
/***********************************************************************\
* Given a starting node (0 for root) and a pathname, return the node *
* for the end item on the path. It MUST BE A DIRECTORY. If the 'CREATE' *
* option is true, then create any missing nodes in the path and create *
* and return the final node as well. *
* This is used to set up a directory, before making nodes in it.. *
* *
* Warning: This function is RECURSIVE. *
* char *orig_path, find this dir (err if not dir) *
* dn_p dirnode, starting point (0 = root) *
* int create, create path if not found *
* dn_p *dn_pp) where to return the node of the dir *
\***********************************************************************/
/*proto*/
int
dev_finddir(char *orig_path, dn_p dirnode, int create, dn_p *dn_pp)
{
devnm_p devnmp;
dn_p dnp;
char pathbuf[DEVMAXPATHSIZE];
char *path;
char *name;
register char *cp;
int retval;
DBPRINT(("dev_finddir\n"));
/***************************************\
* If no parent directory is given *
* then start at the root of the tree *
\***************************************/
if(!dirnode) dirnode = dev_root->dnp;
/***************************************\
* Sanity Checks *
\***************************************/
if(dirnode->type != DEV_DIR) return ENOTDIR;
if(strlen(orig_path) > (DEVMAXPATHSIZE - 1)) return ENAMETOOLONG;
path = pathbuf;
strcpy(path,orig_path);
/***************************************\
* always absolute, skip leading / *
* get rid of / or // or /// etc. *
\***************************************/
while(*path == '/') path++;
/***************************************\
* If nothing left, then parent was it.. *
\***************************************/
if ( *path == '\0' ) {
*dn_pp = dirnode;
return 0;
}
/***************************************\
* find the next segment of the name *
\***************************************/
cp = name = path;
while((*cp != '/') && (*cp != 0)) {
cp++;
}
/***********************************************\
* Check to see if it's the last component *
\***********************************************/
if(*cp) {
path = cp + 1; /* path refers to the rest */
*cp = 0; /* name is now a separate string */
if(!(*path)) {
path = (char *)0; /* was trailing slash */
}
} else {
path = NULL; /* no more to do */
}
/***************************************\
* Start scanning along the linked list *
\***************************************/
devnmp = dev_findname(dirnode,name);
if(devnmp) { /* check it's a directory */
dnp = devnmp->dnp;
if(dnp->type != DEV_DIR) return ENOTDIR;
} else {
/***************************************\
* The required element does not exist *
* So we will add it if asked to. *
\***************************************/
if(!create) return ENOENT;
if(retval = dev_add_entry(name, dirnode, DEV_DIR,
NULL, NULL, NULL, &devnmp)) {
return retval;
}
dnp = devnmp->dnp;
devfs_propogate(dirnode->by.Dir.myname,devnmp);
}
if(path != NULL) { /* decide whether to recurse more or return */
return (dev_finddir(path,dnp,create,dn_pp));
} else {
*dn_pp = dnp;
return 0;
}
}
/***********************************************************************\
* Add a new NAME element to the devfs *
* If we're creating a root node, then dirname is NULL *
* Basically this creates a new namespace entry for the device node *
* *
* Creates a name node, and links it to the supplied node *
\***********************************************************************/
/*proto*/
int
dev_add_name(char *name, dn_p dirnode, devnm_p back, dn_p dnp,
devnm_p *devnm_pp)
{
devnm_p devnmp;
DBPRINT(("dev_add_name\n"));
if(dirnode != NULL ) {
if(dirnode->type != DEV_DIR) return(ENOTDIR);
if( dev_findname(dirnode,name))
return(EEXIST);
}
/*
* make sure the name is legal
* slightly misleading in the case of NULL
*/
if( !name || (strlen(name) > (DEVMAXNAMESIZE - 1)))
return (ENAMETOOLONG);
/*
* Allocate and fill out a new directory entry
*/
if(!(devnmp = (devnm_p)malloc(sizeof(devnm_t),
M_DEVFSNAME, M_NOWAIT))) {
return ENOMEM;
}
bzero(devnmp,sizeof(devnm_t));
/* inherrit our parent's mount info */ /*XXX*/
/* a kludge but.... */
if(dirnode && ( dnp->dvm == NULL)) {
dnp->dvm = dirnode->dvm;
if(!dnp->dvm) printf("parent had null dvm ");
}
/*
* Link the two together
* include the implicit link in the count of links to the devnode..
* this stops it from being accidentally freed later.
*/
devnmp->dnp = dnp;
dnp->links++ ; /* implicit from our own name-node */
/*
* Make sure that we can find all the links that reference a node
* so that we can get them all if we need to zap the node.
*/
if(dnp->linklist) {
devnmp->nextlink = dnp->linklist;
devnmp->prevlinkp = devnmp->nextlink->prevlinkp;
devnmp->nextlink->prevlinkp = &(devnmp->nextlink);
*devnmp->prevlinkp = devnmp;
dnp->linklist = devnmp;
} else {
devnmp->nextlink = devnmp;
devnmp->prevlinkp = &(devnmp->nextlink);
dnp->linklist = devnmp;
}
/*
* If the node is a directory, then we need to handle the
* creation of the .. link.
* A NULL dirnode indicates a root node, so point to ourself.
*/
if(dnp->type == DEV_DIR) {
dnp->by.Dir.myname = devnmp;
/*
* If we are unlinking from an old dir, decrement its links
* as we point our '..' elsewhere
* Note: it's up to the calling code to remove the
* us from the original directory's list
*/
if(dnp->by.Dir.parent) {
dnp->by.Dir.parent->links--;
}
if(dirnode) {
dnp->by.Dir.parent = dirnode;
} else {
dnp->by.Dir.parent = dnp;
}
dnp->by.Dir.parent->links++; /* account for the new '..' */
}
/*
* put the name into the directory entry.
*/
strcpy(devnmp->name, name);
/*
* Check if we are not making a root node..
* (i.e. have parent)
*/
if(dirnode) {
/*
* Put it on the END of the linked list of directory entries
*/
devnmp->parent = dirnode; /* null for root */
devnmp->prevp = dirnode->by.Dir.dirlast;
devnmp->next = *(devnmp->prevp); /* should be NULL */ /*right?*/
*(devnmp->prevp) = devnmp;
dirnode->by.Dir.dirlast = &(devnmp->next);
dirnode->by.Dir.entrycount++;
dirnode->len += strlen(name) + 8;/*ok, ok?*/
}
*devnm_pp = devnmp;
return 0 ;
}
/***********************************************************************\
* Add a new element to the devfs plane. *
* *
* Creates a new dev_node to go with it if the prototype should not be *
* reused. (Is a DIR, or we select SPLIT_DEVS at compile time) *
* 'by' gives us info to make our node if we don't have a prototype. *
* If 'by is null and proto exists, then the 'by' field of *
* the proto is used intead in the CREATE case. *
* note the 'links' count is 0 (except if a dir) *
* but it is only cleared on a transition *
* so this is ok till we link it to something *
* Even in SPLIT_DEVS mode, *
* if the node already exists on the wanted plane, just return it *
\***********************************************************************/
/*proto*/
int
dev_add_node(int entrytype, union typeinfo *by, dn_p proto,
dn_p *dn_pp,struct devfsmount *dvm)
{
dn_p dnp;
DBPRINT(("dev_add_node\n"));
#if defined SPLIT_DEVS
/*
* If we have a prototype, then check if there is already a sibling
* on the mount plane we are looking at, if so, just return it.
*/
if (proto) {
dnp = proto->nextsibling;
while( dnp != proto) {
if (dnp->dvm == dvm) {
*dn_pp = dnp;
return (0);
}
dnp = dnp->nextsibling;
}
if (by == NULL)
by = &(proto->by);
}
#else /* SPLIT_DEVS */
if ( proto ) {
switch (proto->type) {
case DEV_BDEV:
case DEV_CDEV:
case DEV_DDEV:
*dn_pp = proto;
return 0;
}
}
#endif /* SPLIT_DEVS */
if(!(dnp = (dn_p)malloc(sizeof(devnode_t),
M_DEVFSNODE, M_NOWAIT)))
{
return ENOMEM;
}
/*
* If we have a proto, that means that we are duplicating some
* other device, which can only happen if we are not at the back plane
*/
if(proto) {
bcopy(proto, dnp, sizeof(devnode_t));
dnp->links = 0;
dnp->linklist = NULL;
dnp->vn = NULL;
dnp->len = 0;
/* add to END of siblings list */
dnp->prevsiblingp = proto->prevsiblingp;
*(dnp->prevsiblingp) = dnp;
dnp->nextsibling = proto;
proto->prevsiblingp = &(dnp->nextsibling);
} else {
/*
* We have no prototype, so start off with a clean slate
*/
bzero(dnp,sizeof(devnode_t));
dnp->type = entrytype;
getnanotime(&(dnp->ctime));
dnp->mtime = dnp->ctime;
dnp->atime = dnp->ctime;
dnp->nextsibling = dnp;
dnp->prevsiblingp = &(dnp->nextsibling);
}
dnp->dvm = dvm;
/*
* fill out the dev node according to type
*/
switch(entrytype) {
case DEV_DIR:
/*
* As it's a directory, make sure
* it has a null entries list
*/
dnp->by.Dir.dirlast = &(dnp->by.Dir.dirlist);
dnp->by.Dir.dirlist = (devnm_p)0;
dnp->by.Dir.entrycount = 0;
/* until we know better, it has a null parent pointer*/
dnp->by.Dir.parent = NULL;
dnp->links++; /* for .*/
dnp->by.Dir.myname = NULL;
/*
* make sure that the ops associated with it are the ops
* that we use (by default) for directories
*/
dnp->ops = &devfs_vnodeop_p;
dnp->mode |= 0555; /* default perms */
break;
case DEV_SLNK:
/*
* As it's a symlink allocate and store the link info
* Symlinks should only ever be created by the user,
* so they are not on the back plane and should not be
* propogated forward.. a bit like directories in that way..
* A symlink only exists on one plane and has its own
* node.. therefore we might be on any random plane.
*/
dnp->by.Slnk.name = malloc(by->Slnk.namelen+1,
M_DEVFSNODE, M_NOWAIT);
if (!dnp->by.Slnk.name) {
free(dnp,M_DEVFSNODE);
return ENOMEM;
}
strncpy(dnp->by.Slnk.name,by->Slnk.name,by->Slnk.namelen);
dnp->by.Slnk.namelen = by->Slnk.namelen;
dnp->ops = &devfs_vnodeop_p;
dnp->mode |= 0555; /* default perms */
break;
case DEV_BDEV:
/*
* Make sure it has DEVICE type ops
* and device specific fields are correct
*/
dnp->ops = &dev_spec_vnodeop_p;
dnp->by.Bdev.bdevsw = by->Bdev.bdevsw;
dnp->by.Bdev.dev = by->Bdev.dev;
break;
case DEV_CDEV:
/*
* Make sure it has DEVICE type ops
* and device specific fields are correct
*/
dnp->ops = &dev_spec_vnodeop_p;
dnp->by.Cdev.cdevsw = by->Cdev.cdevsw;
dnp->by.Cdev.dev = by->Cdev.dev;
break;
case DEV_DDEV:
/*
* store the address of (the address of) the ops
* and the magic cookie to use with them
*/
dnp->by.Ddev.arg = by->Ddev.arg;
dnp->by.Ddev.ops = by->Ddev.ops;
dnp->ops = by->Ddev.ops;
break;
default:
return EINVAL;
}
*dn_pp = dnp;
return 0 ;
}
/*proto*/
int
dev_touch(devnm_p key) /* update the node for this dev */
{
DBPRINT(("dev_touch\n"));
getnanotime(&(key->dnp->mtime));
return 0; /*XXX*/
}
/*proto*/
void
devfs_dn_free(dn_p dnp)
{
if(--dnp->links <= 0 ) /* can be -1 for initial free, on error */
{
/*probably need to do other cleanups XXX */
if (dnp->nextsibling != dnp) {
dn_p* prevp = dnp->prevsiblingp;
*prevp = dnp->nextsibling;
dnp->nextsibling->prevsiblingp = prevp;
}
if(dnp->type == DEV_SLNK) {
free(dnp->by.Slnk.name,M_DEVFSNODE);
}
devfs_dropvnode(dnp);
free (dnp, M_DEVFSNODE);
}
}
/***********************************************************************\
* Front Node Operations *
* Add or delete a chain of front nodes *
\***********************************************************************/
/***********************************************************************\
* Given a directory backing node, and a child backing node, add the *
* appropriate front nodes to the front nodes of the directory to *
* represent the child node to the user *
* *
* on failure, front nodes will either be correct or not exist for each *
* front dir, however dirs completed will not be stripped of completed *
* frontnodes on failure of a later frontnode *
* *
* This allows a new node to be propogated through all mounted planes *
* *
\***********************************************************************/
/*proto*/
int
devfs_propogate(devnm_p parent,devnm_p child)
{
int error;
devnm_p newnmp;
dn_p dnp = child->dnp;
dn_p pdnp = parent->dnp;
dn_p adnp = parent->dnp;
int type = child->dnp->type;
DBPRINT((" devfs_propogate\n"));
/***********************************************\
* Find the other instances of the parent node *
\***********************************************/
for (adnp = pdnp->nextsibling;
adnp != pdnp;
adnp = adnp->nextsibling)
{
/*
* Make the node, using the original as a prototype)
* if the node already exists on that plane it won't be
* re-made..
*/
if ( error = dev_add_entry(child->name, adnp, type,
NULL, dnp, adnp->dvm, &newnmp)) {
printf("duplicating %s failed\n",child->name);
}
}
return 0; /* for now always succeed */
}
/***********************************************************************
* remove all instances of this devicename [for backing nodes..]
* note.. if there is another link to the node (non dir nodes only)
* then the devfs_node will still exist as the ref count will be non-0
* removing a directory node will remove all sup-nodes on all planes (ZAP)
*
* Used by device drivers to remove nodes that are no longer relevant
* The argument is the 'cookie' they were given when they created the node
* this function is exported.. see sys/devfsext.h
***********************************************************************/
void
devfs_remove_dev(void *devnmp)
{
dn_p dnp = ((devnm_p)devnmp)->dnp;
dn_p dnp2;
DBPRINT(("devfs_remove_dev\n"));
/* keep removing the next sibling till only we exist. */
while((dnp2 = dnp->nextsibling) != dnp) {
/*
* Keep removing the next front node till no more exist
*/
dnp->nextsibling = dnp2->nextsibling;
dnp->nextsibling->prevsiblingp = &(dnp->nextsibling);
dnp2->nextsibling = dnp2;
dnp2->prevsiblingp = &(dnp2->nextsibling);
while(dnp2->linklist)
{
dev_free_name(dnp2->linklist);
}
}
/*
* then free the main node
* If we are not running in SPLIT_DEVS mode, then
* THIS is what gets rid of the propogated nodes.
*/
while(dnp->linklist)
{
dev_free_name(dnp->linklist);
}
return ;
}
/***************************************************************
* duplicate the backing tree into a tree of nodes hung off the
* mount point given as the argument. Do this by
* calling dev_dup_entry which recurses all the way
* up the tree..
* If we are the first plane, just return the base root
**************************************************************/
/*proto*/
int
dev_dup_plane(struct devfsmount *devfs_mp_p)
{
devnm_p new;
int error = 0;
DBPRINT((" dev_dup_plane\n"));
if(devfs_up_and_going) {
if(error = dev_dup_entry(NULL, dev_root, &new, devfs_mp_p)) {
return error;
}
} else { /* we are doing the dummy mount during initialisation.. */
new = dev_root;
}
devfs_mp_p->plane_root = new;
return error;
}
/***************************************************************\
* Free a whole plane
\***************************************************************/
/*proto*/
void
devfs_free_plane(struct devfsmount *devfs_mp_p)
{
devnm_p devnmp;
DBPRINT((" devfs_free_plane\n"));
devnmp = devfs_mp_p->plane_root;
if(devnmp) {
dev_free_hier(devnmp);
dev_free_name(devnmp);
}
devfs_mp_p->plane_root = NULL;
}
/***************************************************************\
* Create and link in a new front element.. *
* Parent can be 0 for a root node *
* Not presently usable to make a symlink XXX *
* (Ok, symlinks don't propogate)
* recursively will create subnodes corresponding to equivalent *
* child nodes in the base level *
\***************************************************************/
/*proto*/
int
dev_dup_entry(dn_p parent, devnm_p back, devnm_p *dnm_pp,
struct devfsmount *dvm)
{
devnm_p newnmp;
devnm_p newback;
devnm_p newfront;
int error;
dn_p dnp = back->dnp;
int type = dnp->type;
DBPRINT((" dev_dup_entry\n"));
/*
* go get the node made (if we need to)
* use the back one as a prototype
*/
if ( error = dev_add_entry(back->name, parent, type,
NULL, dnp,
parent?parent->dvm:dvm, &newnmp)) {
printf("duplicating %s failed\n",back->name);
}
/*
* If we have just made the root, then insert the pointer to the
* mount information
*/
if(dvm) {
newnmp->dnp->dvm = dvm;
}
/*
* If it is a directory, then recurse down all the other
* subnodes in it....
* note that this time we don't pass on the mount info..
*/
if (type == DEV_DIR)
{
for(newback = back->dnp->by.Dir.dirlist;
newback; newback = newback->next)
{
if(error = dev_dup_entry(newnmp->dnp,
newback, &newfront, NULL))
{
break; /* back out with an error */
}
}
}
*dnm_pp = newnmp;
return error;
}
/***************************************************************\
* Free a name node *
* remember that if there are other names pointing to the *
* dev_node then it may not get freed yet *
* can handle if there is no dnp *
\***************************************************************/
/*proto*/
int
dev_free_name(devnm_p devnmp)
{
dn_p parent = devnmp->parent;
dn_p dnp = devnmp->dnp;
DBPRINT((" dev_free_name\n"));
if(dnp) {
if(dnp->type == DEV_DIR)
{
if(dnp->by.Dir.dirlist)
return (ENOTEMPTY);
devfs_dn_free(dnp); /* account for '.' */
devfs_dn_free(dnp->by.Dir.parent); /* '..' */
}
/*
* unlink us from the list of links for this node
* If we are the only link, it's easy!
* if we are a DIR of course there should not be any
* other links.
*/
if(devnmp->nextlink == devnmp) {
dnp->linklist = NULL;
} else {
if(dnp->linklist == devnmp) {
dnp->linklist = devnmp->nextlink;
}
devnmp->nextlink->prevlinkp = devnmp->prevlinkp;
*devnmp->prevlinkp = devnmp->nextlink;
}
devfs_dn_free(dnp);
}
/*
* unlink ourselves from the directory on this plane
*/
if(parent) /* if not fs root */
{
if( (*devnmp->prevp = devnmp->next) )/* yes, assign */
{
devnmp->next->prevp = devnmp->prevp;
}
else
{
parent->by.Dir.dirlast
= devnmp->prevp;
}
parent->by.Dir.entrycount--;
parent->len -= strlen(devnmp->name) + 8;
}
/***************************************************************\
* If the front node has its own devnode structure, *
* then free it. *
\***************************************************************/
free(devnmp,M_DEVFSNAME);
return 0;
}
/***************************************************************\
* Free a hierarchy starting at a directory node name *
* remember that if there are other names pointing to the *
* dev_node then it may not get freed yet *
* can handle if there is no dnp *
* leave the node itself allocated. *
\***************************************************************/
/*proto*/
void
dev_free_hier(devnm_p devnmp)
{
dn_p dnp = devnmp->dnp;
DBPRINT((" dev_free_hier\n"));
if(dnp) {
if(dnp->type == DEV_DIR)
{
while(dnp->by.Dir.dirlist)
{
dev_free_hier(dnp->by.Dir.dirlist);
dev_free_name(dnp->by.Dir.dirlist);
}
}
}
}
/*******************************************************\
*********************************************************
* ROUTINES to control the connection between devfs *
* nodes and the system's vnodes *
*********************************************************
\*******************************************************/
/*******************************************************\
* Theoretically this could be called for any kind of *
* vnode, however in practice it must be a DEVFS vnode *
\*******************************************************/
/*proto*/
int
devfs_vntodn(struct vnode *vn_p, dn_p *dn_pp)
{
DBPRINT((" vntodn "));
if(vn_p->v_tag != VT_DEVFS)
{
printf("bad-tag2 ");
Debugger("bad-tag ");
return(EINVAL);
}
#if 0
/*
* XXX: This is actually a "normal" case when vclean calls us without
* XXX: incrementing the reference count first.
*/
if(vn_p->v_usecount == 0)
{
printf("No references! ");
}
#endif
switch(vn_p->v_type) {
case VBAD:
printf("bad-type2 (VBAD)");
return(EINVAL);
#if 0
case VNON:
printf("bad-type2 (VNON)");
return(EINVAL);
#endif
}
*dn_pp = (dn_p)vn_p->v_data;
return(0);
}
/***************************************************************\
* given a dev_node, find the appropriate vnode if one is already*
* associated, or get a new one an associate it with the dev_node*
* need to check about vnode references.. should we increment it?*
\***************************************************************/
/*proto*/
int
devfs_dntovn(dn_p dnp, struct vnode **vn_pp)
{
struct vnode *vn_p, *nvp;
int error = 0;
struct proc *p = curproc; /* XXX */
vn_p = dnp->vn;
DBPRINT(("dntovn "));
if( vn_p)
{
if(vn_p->v_id != dnp->vn_id)
{
#if 0
/* XXX: This is `normal'... */
printf("bad-id ");
#endif
goto skip;
}
if(vn_p->v_tag != VT_DEVFS)
{
#if 0
/* XXX: This is `normal'... */
printf("bad-tag ");
#endif
goto skip;
}
if(vn_p->v_op != *(dnp->ops))
{
printf("bad-ops ");
goto skip;
}
if((dn_p)(vn_p->v_data) != dnp)
{
printf("bad-rev_link ");
goto skip;
}
if(vn_p->v_type != VNON)
{
vget(vn_p, LK_EXCLUSIVE, p);
*vn_pp = vn_p;
return(0);
}
else
{
printf("bad-type");
}
skip:
vn_p = (struct vnode *) 0;
}
if(!(error = getnewvnode(VT_DEVFS,
dnp->dvm->mount,
*(dnp->ops),
&vn_p)))
{
dnp->vn = vn_p;
dnp->vn_id = vn_p->v_id;
*vn_pp = vn_p;
DBPRINT(("(New vnode)"));
switch(dnp->type)
{
case DEV_SLNK:
vn_p->v_type = VLNK;
break;
case DEV_DIR:
if(dnp->by.Dir.parent == dnp)
{
vn_p->v_flag |= VROOT;
}
vn_p->v_type = VDIR;
break;
case DEV_BDEV:
vn_p->v_type = VBLK;
if (nvp = checkalias(vn_p,
dnp->by.Bdev.dev,
(struct mount *)0))
{
vput(vn_p);
vn_p = nvp;
}
break;
case DEV_CDEV:
vn_p->v_type = VCHR;
if (nvp = checkalias(vn_p,
dnp->by.Cdev.dev,
(struct mount *)0))
{
vput(vn_p);
vn_p = nvp;
}
break;
case DEV_DDEV:
break;
}
if ( vn_p)
{
vn_p->v_mount = dnp->dvm->mount;/* XXX Duplicated */
*vn_pp = vn_p;
vn_p->v_data = (void *)dnp;
}
else
{
error = EINVAL;
}
vn_lock(vn_p, LK_EXCLUSIVE | LK_RETRY, p);
}
return error;
}
/***********************************************************************\
* add a whole device, with no prototype.. make name element and node *
* Used for adding the original device entries *
\***********************************************************************/
/*proto*/
int
dev_add_entry(char *name, dn_p parent, int type, union typeinfo *by,
dn_p proto, struct devfsmount *dvm, devnm_p *nm_pp)
{
dn_p dnp;
int error = 0;
DBPRINT((" devfs_add_entry\n"));
if (error = dev_add_node(type, by, proto, &dnp,
(parent?parent->dvm:dvm)))
{
printf("Device %s: base node allocation failed (Errno=%d)\n",
name,error);
return error;
}
if ( error = dev_add_name(name ,parent ,NULL, dnp, nm_pp))
{
devfs_dn_free(dnp); /* 1->0 for dir, 0->(-1) for other */
printf("Device %s: name slot allocation failed (Errno=%d)\n",
name,error);
}
return error;
}
/***********************************************************************\
* Add the named device entry into the given directory, and make it *
* The appropriate type... (called (sometimes indirectly) by drivers..) *
* this function is exported.. see sys/devfsext.h *
* Has the capacity to take printf type arguments to format the device *
* names *
\***********************************************************************/
void *
devfs_add_devswf(void *devsw, int minor, int chrblk, uid_t uid,
gid_t gid, int perms, char *fmt, ...)
{
int major;
devnm_p new_dev;
dn_p dnp; /* devnode for parent directory */
struct cdevsw *cd;
struct bdevsw *bd;
int retval;
union typeinfo by;
va_list ap;
char *name, *path, buf[256]; /* XXX */
int i;
va_start(ap, fmt);
i = kvprintf(fmt, NULL, (void*)buf, 32, ap);
va_end(ap);
buf[i] = '\0';
name = NULL;
for(i=strlen(buf); i>0; i--)
if(buf[i] == '/') {
name=&buf[i];
buf[i]=0;
break;
}
if (name) {
*name++ = '\0';
path = buf;
} else {
name = buf;
path = "/";
}
DBPRINT(("dev_add\n"));
retval = dev_finddir(path,NULL,1,&dnp);
if (retval) return 0;
switch(chrblk)
{
case DV_CHR:
cd = devsw;
major = cd->d_maj;
if ( major == -1 ) return NULL;
by.Cdev.cdevsw = cd;
by.Cdev.dev = makedev(major, minor);
if( dev_add_entry(name, dnp, DEV_CDEV, &by, NULL, NULL, &new_dev))
return NULL;
break;
case DV_BLK:
bd = devsw;
major = bd->d_maj;
if ( major == -1 ) return NULL;
by.Bdev.bdevsw = bd;
by.Bdev.dev = makedev(major, minor);
if( dev_add_entry(name, dnp, DEV_BDEV, &by, NULL, NULL, &new_dev))
return NULL;
break;
default:
return NULL;
}
new_dev->dnp->gid = gid;
new_dev->dnp->uid = uid;
new_dev->dnp->mode |= perms;
devfs_propogate(dnp->by.Dir.myname,new_dev);
return new_dev;
}
/***********************************************************************\
* Add the named device entry into the given directory, and make it *
* a link to the already created device given as an arg.. *
* this function is exported.. see sys/devfsext.h *
\***********************************************************************/
void *
devfs_link(void *original, char *fmt, ...)
{
devnm_p new_dev;
devnm_p orig = (devnm_p) original;
dn_p dirnode; /* devnode for parent directory */
int retval;
va_list ap;
char *p, buf[256]; /* XXX */
int i;
va_start(ap, fmt);
i = kvprintf(fmt, NULL, (void*)buf, 32, ap);
va_end(ap);
buf[i] = '\0';
p = NULL;
for(i=strlen(buf); i>0; i--)
if(buf[i] == '/') {
p=&buf[i];
buf[i]=0;
break;
}
DBPRINT(("dev_add\n"));
/*
* The DEV_CDEV below is not used other than it must NOT be DEV_DIR
* the correctness of original should be checked..
*/
if (p) {
*p++ = '\0';
retval = dev_finddir(buf,NULL,1,&dirnode);
if (retval) return 0;
if( dev_add_name(p, dirnode, NULL, orig->dnp, &new_dev))
return NULL;
} else {
retval = dev_finddir("/",NULL,1,&dirnode);
if (retval) return 0;
if( dev_add_name(buf, dirnode, NULL, orig->dnp, &new_dev))
return NULL;
}
devfs_propogate(dirnode->by.Dir.myname,new_dev);
return new_dev;
}
/*
* internal kernel call to open a device. Return either 0 or an open vnode.
*/
struct vnode *
devfs_open_device(char *path, int type)
{
register char *lastslash;
char *nextpart;
devnm_p nm_p;
dn_p dirnode;
struct vnode *vn;
/*
* If the caller didn't supply a full path, ignore and be
* noisy about it.
*/
if (*path != '/') {
printf (__FUNCTION__ ": caller supplied bad path\n");
return (NULL);
}
/*
* find the last '/'. Unfortunatly rindex() while being in
* libkern source, is not being compiled.. do it by hand.
* lastslash = strrchr(path,(int)'c');
* There will be at LEAST one '/'.
*/
{
register char *p = path; /* don't destroy path */
for (lastslash = NULL;*p; ++p) {
if (*p == '/')
lastslash = p;
}
}
dirnode = dev_root->dnp;
if(lastslash != path) {
/* find the directory we need */
*lastslash = '\0';
if (dev_finddir(path, dirnode, NULL, &dirnode) != 0) {
*lastslash = '/';
return (NULL);
}
/* ok we found the directory, put the slash back */
*lastslash = '/';
}
nextpart = ++lastslash;
if (*nextpart == '\0')
return (NULL);
/*
* Now only return true if it exists and is the right type.
*/
if ((nm_p = dev_findname(dirnode, nextpart)) == NULL) {
return (NULL);
}
switch(type) {
case DV_BLK:
if( nm_p->dnp->type != DEV_BDEV)
return (NULL);
break;
case DV_CHR:
if( nm_p->dnp->type != DEV_CDEV)
return (NULL);
break;
}
if ( devfs_dntovn(nm_p->dnp, &vn))
return (NULL);
#if 0
if ( VOP_OPEN(vn, FREAD, proc0.p_cred->pc_ucred, &proc0)) {
vput(vn);
return (NULL);
}
#endif
return (vn);
}
/*
* internal kernel call to close a devfs device.
* It should have been openned by th ecall above.
* try not mix it with user-openned vnodes.
* Frees the vnode.
*/
void
devfs_close_device(struct vnode *vn)
{
#if 0
VOP_CLOSE(vn, 0, proc0.p_cred->pc_ucred, &proc0) ;
#endif
vput(vn);
}
/*
* Little utility routine for compatibilty.
* Returns the dev_t that a devfs vnode represents.
* should go away after dev_t go away :).
*/
dev_t
devfs_vntodev(struct vnode *vn)
{
register dn_p dnp;
dnp = (dn_p)vn->v_data;
switch (dnp->type) {
case DEV_BDEV:
return (dnp->by.Bdev.dev);
break;
case DEV_CDEV:
return (dnp->by.Cdev.dev);
break;
}
panic ("bad devfs DEVICE vnode");
}