freebsd-nq/sys/kern/kern_lkm.c
1995-05-30 08:16:23 +00:00

928 lines
20 KiB
C

/*-
* Copyright (c) 1992 Terrence R. Lambert.
* Copyright (c) 1994 Christopher G. Demetriou
* 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 Terrence R. Lambert.
* 4. The name Terrence R. Lambert may not be used to endorse or promote
* products derived from this software without specific prior written
* permission.
*
* THIS SOFTWARE IS PROVIDED BY TERRENCE R. LAMBERT ``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 TERRENCE R. LAMBERT 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: kern_lkm.c,v 1.12 1995/04/20 05:08:31 wpaul Exp $
*/
/*
* XXX it's not really safe to unload *any* of the types which are
* currently loadable; e.g. you could unload a syscall which was being
* blocked in, etc. In the long term, a solution should be come up
* with, but "not right now." -- cgd
*/
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/ioctl.h>
#include <sys/tty.h>
#include <sys/conf.h>
#include <sys/file.h>
#include <sys/proc.h>
#include <sys/uio.h>
#include <sys/kernel.h>
#include <sys/malloc.h>
#include <sys/mount.h>
#include <sys/sysent.h>
#include <sys/exec.h>
#include <sys/imgact.h>
#include <sys/lkm.h>
#include <vm/vm.h>
#include <vm/vm_param.h>
#include <vm/vm_kern.h>
#define PAGESIZE 1024 /* kmem_alloc() allocation quantum */
#define LKM_ALLOC 0x01
#define LKM_WANT 0x02
#define LKMS_IDLE 0x00
#define LKMS_RESERVED 0x01
#define LKMS_LOADING 0x02
#define LKMS_LOADED 0x04
#define LKMS_UNLOADING 0x08
static int lkm_v = 0;
static int lkm_state = LKMS_IDLE;
#ifndef MAXLKMS
#define MAXLKMS 20
#endif
static struct lkm_table lkmods[MAXLKMS]; /* table of loaded modules */
static struct lkm_table *curp; /* global for in-progress ops */
/*ARGSUSED*/
int
lkmcopen(dev, flag, devtype, p)
dev_t dev;
int flag;
int devtype;
struct proc *p;
{
int error;
if (minor(dev) != 0)
return(ENXIO); /* bad minor # */
/*
* Use of the loadable kernel module device must be exclusive; we
* may try to remove this restriction later, but it's really no
* hardship.
*/
while (lkm_v & LKM_ALLOC) {
if (flag & FNONBLOCK) /* don't hang */
return(EBUSY);
lkm_v |= LKM_WANT;
/*
* Sleep pending unlock; we use tsleep() to allow
* an alarm out of the open.
*/
error = tsleep((caddr_t)&lkm_v, TTIPRI|PCATCH, "lkmopn", 0);
if (error)
return(error); /* leave LKM_WANT set -- no problem */
}
lkm_v |= LKM_ALLOC;
return(0); /* pseudo-device open */
}
/*
* Unreserve the memory associated with the current loaded module; done on
* a coerced close of the lkm device (close on premature exit of modload)
* or explicitly by modload as a result of a link failure.
*/
static void
lkmunreserve()
{
if (lkm_state == LKMS_IDLE)
return;
/*
* Actually unreserve the memory
*/
if (curp && curp->area) {
kmem_free(kmem_map, curp->area, curp->size);/**/
curp->area = 0;
if (curp->private.lkm_any != NULL)
curp->private.lkm_any = NULL;
}
lkm_state = LKMS_IDLE;
}
int
lkmcclose(dev, flag, mode, p)
dev_t dev;
int flag;
int mode;
struct proc *p;
{
if (!(lkm_v & LKM_ALLOC)) {
#ifdef DEBUG
printf("LKM: close before open!\n");
#endif /* DEBUG */
return(EBADF);
}
/* do this before waking the herd... */
if (curp && !curp->used) {
/*
* If we close before setting used, we have aborted
* by way of error or by way of close-on-exit from
* a premature exit of "modload".
*/
lkmunreserve(); /* coerce state to LKM_IDLE */
}
lkm_v &= ~LKM_ALLOC;
wakeup((caddr_t)&lkm_v); /* thundering herd "problem" here */
return(0); /* pseudo-device closed */
}
/*ARGSUSED*/
int
lkmcioctl(dev, cmd, data, flag)
dev_t dev;
int cmd;
caddr_t data;
int flag;
{
int err = 0;
int i;
struct lmc_resrv *resrvp;
struct lmc_loadbuf *loadbufp;
struct lmc_unload *unloadp;
struct lmc_stat *statp;
char istr[MAXLKMNAME];
switch(cmd) {
case LMRESERV: /* reserve pages for a module */
if ((flag & FWRITE) == 0) /* only allow this if writing */
return EPERM;
resrvp = (struct lmc_resrv *)data;
/*
* Find a free slot.
*/
for (i = 0; i < MAXLKMS; i++)
if (!lkmods[i].used)
break;
if (i == MAXLKMS) {
err = ENOMEM; /* no slots available */
break;
}
curp = &lkmods[i];
curp->id = i; /* self reference slot offset */
resrvp->slot = i; /* return slot */
/*
* Get memory for module
*/
curp->size = resrvp->size;
curp->area = kmem_alloc(kmem_map, curp->size);/**/
curp->offset = 0; /* load offset */
resrvp->addr = curp->area; /* ret kernel addr */
#ifdef DEBUG
printf("LKM: LMRESERV (actual = 0x%08x)\n", curp->area);
printf("LKM: LMRESERV (adjusted = 0x%08x)\n",
trunc_page(curp->area));
#endif /* DEBUG */
lkm_state = LKMS_RESERVED;
break;
case LMLOADBUF: /* Copy in; stateful, follows LMRESERV */
if ((flag & FWRITE) == 0) /* only allow this if writing */
return EPERM;
loadbufp = (struct lmc_loadbuf *)data;
i = loadbufp->cnt;
if ((lkm_state != LKMS_RESERVED && lkm_state != LKMS_LOADING)
|| i < 0
|| i > MODIOBUF
|| i > curp->size - curp->offset) {
err = ENOMEM;
break;
}
/* copy in buffer full of data */
err = copyin((caddr_t)loadbufp->data,
(caddr_t)curp->area + curp->offset, i);
if (err)
break;
if ((curp->offset + i) < curp->size) {
lkm_state = LKMS_LOADING;
#ifdef DEBUG
printf("LKM: LMLOADBUF (loading @ %d of %d, i = %d)\n",
curp->offset, curp->size, i);
#endif /* DEBUG */
} else {
lkm_state = LKMS_LOADED;
#ifdef DEBUG
printf("LKM: LMLOADBUF (loaded)\n");
#endif /* DEBUG */
}
curp->offset += i;
break;
case LMUNRESRV: /* discard reserved pages for a module */
if ((flag & FWRITE) == 0) /* only allow this if writing */
return EPERM;
lkmunreserve(); /* coerce state to LKM_IDLE */
#ifdef DEBUG
printf("LKM: LMUNRESERV\n");
#endif /* DEBUG */
break;
case LMREADY: /* module loaded: call entry */
if ((flag & FWRITE) == 0) /* only allow this if writing */
return EPERM;
switch (lkm_state) {
case LKMS_LOADED:
break;
case LKMS_LOADING:
/* The remainder must be bss, so we clear it */
bzero((caddr_t)curp->area + curp->offset,
curp->size - curp->offset);
break;
default:
#ifdef DEBUG
printf("lkm_state is %02x\n", lkm_state);
#endif /* DEBUG */
return ENXIO;
}
curp->entry = (int (*)()) (*((int *) (data)));
/* call entry(load)... (assigns "private" portion) */
err = (*(curp->entry))(curp, LKM_E_LOAD, LKM_VERSION);
if (err) {
/*
* Module may refuse loading or may have a
* version mismatch...
*/
lkm_state = LKMS_UNLOADING; /* for lkmunreserve */
lkmunreserve(); /* free memory */
curp->used = 0; /* free slot */
break;
}
/*
* It's possible for a user to load a module that doesn't
* initialize itself correctly. (You can even get away with
* using it for a while.) Unfortunately, we are faced with
* the following problems:
* - we can't tell a good module from a bad one until
* after we've run its entry function (if the private
* section is uninitalized after we return from the
* entry, then something's fishy)
* - now that we've called the entry function, we can't
* forcibly unload the module without risking a crash
* - since we don't know what the module's entry function
* did, we can't easily clean up the mess it may have
* made, so we can't know just how unstable the system
* may be
* So, being stuck between a rock and a hard place, we
* have no choice but to do this...
*/
if (curp->private.lkm_any == NULL)
panic("loadable module initialization failed");
curp->used = 1;
#ifdef DEBUG
printf("LKM: LMREADY\n");
#endif /* DEBUG */
lkm_state = LKMS_IDLE;
break;
case LMUNLOAD: /* unload a module */
if ((flag & FWRITE) == 0) /* only allow this if writing */
return EPERM;
unloadp = (struct lmc_unload *)data;
if ((i = unloadp->id) == -1) { /* unload by name */
/*
* Copy name and lookup id from all loaded
* modules. May fail.
*/
err =copyinstr(unloadp->name, istr, MAXLKMNAME-1, NULL);
if (err)
break;
/*
* look up id...
*/
for (i = 0; i < MAXLKMS; i++) {
if (!lkmods[i].used)
continue;
if (!strcmp(istr,
lkmods[i].private.lkm_any->lkm_name))
break;
}
}
/*
* Range check the value; on failure, return EINVAL
*/
if (i < 0 || i >= MAXLKMS) {
err = EINVAL;
break;
}
curp = &lkmods[i];
if (!curp->used) {
err = ENOENT;
break;
}
/* call entry(unload) */
if ((*(curp->entry))(curp, LKM_E_UNLOAD, LKM_VERSION)) {
err = EBUSY;
break;
}
lkm_state = LKMS_UNLOADING; /* non-idle for lkmunreserve */
lkmunreserve(); /* free memory */
curp->used = 0; /* free slot */
break;
case LMSTAT: /* stat a module by id/name */
/* allow readers and writers to stat */
statp = (struct lmc_stat *)data;
if ((i = statp->id) == -1) { /* stat by name */
/*
* Copy name and lookup id from all loaded
* modules.
*/
copystr(statp->name, istr, MAXLKMNAME-1, NULL);
/*
* look up id...
*/
for (i = 0; i < MAXLKMS; i++) {
if (!lkmods[i].used)
continue;
if (!strcmp(istr,
lkmods[i].private.lkm_any->lkm_name))
break;
}
if (i == MAXLKMS) { /* Not found */
err = ENOENT;
break;
}
}
/*
* Range check the value; on failure, return EINVAL
*/
if (i < 0 || i >= MAXLKMS) {
err = EINVAL;
break;
}
curp = &lkmods[i];
if (!curp->used) { /* Not found */
err = ENOENT;
break;
}
/*
* Copy out stat information for this module...
*/
statp->id = curp->id;
statp->offset = curp->private.lkm_any->lkm_offset;
statp->type = curp->private.lkm_any->lkm_type;
statp->area = curp->area;
statp->size = curp->size / PAGESIZE;
statp->private = (unsigned long)curp->private.lkm_any;
statp->ver = curp->private.lkm_any->lkm_ver;
copystr(curp->private.lkm_any->lkm_name,
statp->name,
MAXLKMNAME - 2,
NULL);
break;
default: /* bad ioctl()... */
err = ENOTTY;
break;
}
return (err);
}
/*
* Acts like "nosys" but can be identified in sysent for dynamic call
* number assignment for a limited number of calls.
*
* Place holder for system call slots reserved for loadable modules.
*/
int
lkmnosys()
{
return(nosys());
}
/*
* Acts like "enodev", but can be identified in cdevsw and bdevsw for
* dynamic driver major number assignment for a limited number of
* drivers.
*
* Place holder for device switch slots reserved for loadable modules.
*/
int
lkmenodev()
{
return(enodev());
}
int
lkmexists(lkmtp)
struct lkm_table *lkmtp;
{
int i;
/*
* see if name exists...
*/
for (i = 0; i < MAXLKMS; i++) {
/*
* An unused module and the one we are testing are not
* considered.
*/
if (!lkmods[i].used || &lkmods[i] == lkmtp)
continue;
if (!strcmp(lkmtp->private.lkm_any->lkm_name,
lkmods[i].private.lkm_any->lkm_name))
return(1); /* already loaded... */
}
return(0); /* module not loaded... */
}
/*
* For the loadable system call described by the structure pointed to
* by lkmtp, load/unload/stat it depending on the cmd requested.
*/
static int
_lkm_syscall(lkmtp, cmd)
struct lkm_table *lkmtp;
int cmd;
{
struct lkm_syscall *args = lkmtp->private.lkm_syscall;
int i;
int err = 0;
switch(cmd) {
case LKM_E_LOAD:
/* don't load twice! */
if (lkmexists(lkmtp))
return(EEXIST);
if ((i = args->lkm_offset) == -1) { /* auto */
/*
* Search the table looking for a slot...
*/
for (i = 0; i < aout_sysvec.sv_size; i++)
if (aout_sysvec.sv_table[i].sy_call == lkmnosys)
break; /* found it! */
/* out of allocable slots? */
if (i == aout_sysvec.sv_size) {
err = ENFILE;
break;
}
} else { /* assign */
if (i < 0 || i >= aout_sysvec.sv_size) {
err = EINVAL;
break;
}
}
/* save old */
bcopy(&aout_sysvec.sv_table[i],
&(args->lkm_oldent),
sizeof(struct sysent));
/* replace with new */
bcopy(args->lkm_sysent,
&aout_sysvec.sv_table[i],
sizeof(struct sysent));
/* done! */
args->lkm_offset = i; /* slot in sysent[] */
break;
case LKM_E_UNLOAD:
/* current slot... */
i = args->lkm_offset;
/* replace current slot contents with old contents */
bcopy(&(args->lkm_oldent),
&aout_sysvec.sv_table[i],
sizeof(struct sysent));
break;
case LKM_E_STAT: /* no special handling... */
break;
}
return(err);
}
/*
* For the loadable virtual file system described by the structure pointed
* to by lkmtp, load/unload/stat it depending on the cmd requested.
*/
static int
_lkm_vfs(lkmtp, cmd)
struct lkm_table *lkmtp;
int cmd;
{
struct lkm_vfs *args = lkmtp->private.lkm_vfs;
struct vfsconf *vfc = args->lkm_vfsconf;
int i;
int err = 0;
switch(cmd) {
case LKM_E_LOAD:
/* don't load twice! */
if (lkmexists(lkmtp))
return(EEXIST);
for(i = 0; i < MOUNT_MAXTYPE; i++) {
if(!strcmp(vfc->vfc_name, vfsconf[i]->vfc_name)) {
return EEXIST;
}
}
i = args->lkm_offset = vfc->vfc_index;
if (i < 0) {
for (i = MOUNT_MAXTYPE - 1; i >= 0; i--) {
if(vfsconf[i] == &void_vfsconf)
break;
}
}
if (i < 0) {
return EINVAL;
}
args->lkm_offset = vfc->vfc_index = i;
vfsconf[i] = vfc;
vfssw[i] = vfc->vfc_vfsops;
/* like in vfs_op_init */
for(i = 0; args->lkm_vnodeops->ls_items[i]; i++) {
struct vnodeopv_desc *opv =
(struct vnodeopv_desc *)args->lkm_vnodeops->ls_items[i];
*(opv->opv_desc_vector_p) = NULL;
}
vfs_opv_init((struct vnodeopv_desc **)args->lkm_vnodeops->ls_items);
/*
* Call init function for this VFS...
*/
(*(vfssw[vfc->vfc_index]->vfs_init))();
/* done! */
break;
case LKM_E_UNLOAD:
/* current slot... */
i = args->lkm_offset;
if (vfsconf[i]->vfc_refcount) {
return EBUSY;
}
/* replace current slot contents with old contents */
vfssw[i] = (struct vfsops *)0;
vfsconf[i] = &void_vfsconf;
break;
case LKM_E_STAT: /* no special handling... */
break;
}
return(err);
}
/*
* For the loadable device driver described by the structure pointed to
* by lkmtp, load/unload/stat it depending on the cmd requested.
*/
static int
_lkm_dev(lkmtp, cmd)
struct lkm_table *lkmtp;
int cmd;
{
struct lkm_dev *args = lkmtp->private.lkm_dev;
int i;
int err = 0;
switch(cmd) {
case LKM_E_LOAD:
/* don't load twice! */
if (lkmexists(lkmtp))
return(EEXIST);
switch(args->lkm_devtype) {
case LM_DT_BLOCK:
if ((i = args->lkm_offset) == -1) { /* auto */
/*
* Search the table looking for a slot...
*/
for (i = 0; i < nblkdev; i++)
if (bdevsw[i].d_open == lkmenodev)
break; /* found it! */
/* out of allocable slots? */
if (i == nblkdev) {
err = ENFILE;
break;
}
} else { /* assign */
if (i < 0 || i >= nblkdev) {
err = EINVAL;
break;
}
}
/* save old */
bcopy(&bdevsw[i], &(args->lkm_olddev.bdev), sizeof(struct bdevsw));
/* replace with new */
bcopy(args->lkm_dev.bdev, &bdevsw[i], sizeof(struct bdevsw));
/* done! */
args->lkm_offset = i; /* slot in bdevsw[] */
break;
case LM_DT_CHAR:
if ((i = args->lkm_offset) == -1) { /* auto */
/*
* Search the table looking for a slot...
*/
for (i = 0; i < nchrdev; i++)
if (cdevsw[i].d_open == lkmenodev)
break; /* found it! */
/* out of allocable slots? */
if (i == nchrdev) {
err = ENFILE;
break;
}
} else { /* assign */
if (i < 0 || i >= nchrdev) {
err = EINVAL;
break;
}
}
/* save old */
bcopy(&cdevsw[i], &(args->lkm_olddev.cdev), sizeof(struct cdevsw));
/* replace with new */
bcopy(args->lkm_dev.cdev, &cdevsw[i], sizeof(struct cdevsw));
/* done! */
args->lkm_offset = i; /* slot in cdevsw[] */
break;
default:
err = ENODEV;
break;
}
break;
case LKM_E_UNLOAD:
/* current slot... */
i = args->lkm_offset;
switch(args->lkm_devtype) {
case LM_DT_BLOCK:
/* replace current slot contents with old contents */
bcopy(&(args->lkm_olddev.bdev), &bdevsw[i], sizeof(struct bdevsw));
break;
case LM_DT_CHAR:
/* replace current slot contents with old contents */
bcopy(&(args->lkm_olddev.cdev), &cdevsw[i], sizeof(struct cdevsw));
break;
default:
err = ENODEV;
break;
}
break;
case LKM_E_STAT: /* no special handling... */
break;
}
return(err);
}
#ifdef STREAMS
/*
* For the loadable streams module described by the structure pointed to
* by lkmtp, load/unload/stat it depending on the cmd requested.
*/
static int
_lkm_strmod(lkmtp, cmd)
struct lkm_table *lkmtp;
int cmd;
{
struct lkm_strmod *args = lkmtp->private.lkm_strmod;
int i;
int err = 0;
switch(cmd) {
case LKM_E_LOAD:
/* don't load twice! */
if (lkmexists(lkmtp))
return(EEXIST);
break;
case LKM_E_UNLOAD:
break;
case LKM_E_STAT: /* no special handling... */
break;
}
return(err);
}
#endif /* STREAMS */
/*
* For the loadable execution class described by the structure pointed to
* by lkmtp, load/unload/stat it depending on the cmd requested.
*/
static int
_lkm_exec(lkmtp, cmd)
struct lkm_table *lkmtp;
int cmd;
{
struct lkm_exec *args = lkmtp->private.lkm_exec;
int i;
int err = 0;
const struct execsw **execsw =
(const struct execsw **)&execsw_set.ls_items[0];
#if 1
switch(cmd) {
case LKM_E_LOAD:
/* don't load twice! */
if (lkmexists(lkmtp))
return(EEXIST);
if ((i = args->lkm_offset) == -1) { /* auto */
/*
* Search the table looking for a slot...
*/
for (i = 0; execsw[i] != NULL; i++)
if (execsw[i]->ex_imgact == NULL)
break; /* found it! */
/* out of allocable slots? */
if (execsw[i] == NULL) {
err = ENFILE;
break;
}
} else { /* assign */
err = EINVAL;
break;
}
/* save old */
bcopy(&execsw[i], &(args->lkm_oldexec), sizeof(struct execsw*));
/* replace with new */
bcopy(&(args->lkm_exec), &execsw[i], sizeof(struct execsw*));
/* done! */
args->lkm_offset = i; /* slot in execsw[] */
break;
case LKM_E_UNLOAD:
/* current slot... */
i = args->lkm_offset;
/* replace current slot contents with old contents */
bcopy(&(args->lkm_oldexec), &execsw[i], sizeof(struct execsw*));
break;
case LKM_E_STAT: /* no special handling... */
break;
}
#else
err = EINVAL;
#endif
return(err);
}
static const struct execsw lkm_exec_dummy = { NULL, "lkm" };
TEXT_SET(execsw_set, lkm_exec_dummy);
TEXT_SET(execsw_set, lkm_exec_dummy);
TEXT_SET(execsw_set, lkm_exec_dummy);
TEXT_SET(execsw_set, lkm_exec_dummy);
/*
* This code handles the per-module type "wiring-in" of loadable modules
* into existing kernel tables. For "LM_MISC" modules, wiring and unwiring
* is assumed to be done in their entry routines internal to the module
* itself.
*/
int
lkmdispatch(lkmtp, cmd)
struct lkm_table *lkmtp;
int cmd;
{
int err = 0; /* default = success */
switch(lkmtp->private.lkm_any->lkm_type) {
case LM_SYSCALL:
err = _lkm_syscall(lkmtp, cmd);
break;
case LM_VFS:
err = _lkm_vfs(lkmtp, cmd);
break;
case LM_DEV:
err = _lkm_dev(lkmtp, cmd);
break;
#ifdef STREAMS
case LM_STRMOD:
{
struct lkm_strmod *args = lkmtp->private.lkm_strmod;
}
break;
#endif /* STREAMS */
case LM_EXEC:
err = _lkm_exec(lkmtp, cmd);
break;
case LM_MISC: /* ignore content -- no "misc-specific" procedure */
if (lkmexists(lkmtp))
err = EEXIST;
break;
default:
err = ENXIO; /* unknown type */
break;
}
return(err);
}