linux-stable/drivers/sbus/char/cpwatchdog.c

855 lines
21 KiB
C
Raw Normal View History

/* cpwatchdog.c - driver implementation for hardware watchdog
* timers found on Sun Microsystems CP1400 and CP1500 boards.
*
* This device supports both the generic Linux watchdog
* interface and Solaris-compatible ioctls as best it is
* able.
*
* NOTE: CP1400 systems appear to have a defective intr_mask
* register on the PLD, preventing the disabling of
* timer interrupts. We use a timer to periodically
* reset 'stopped' watchdogs on affected platforms.
*
* Copyright (c) 2000 Eric Brower (ebrower@usa.net)
*/
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/fs.h>
#include <linux/errno.h>
#include <linux/major.h>
#include <linux/init.h>
#include <linux/miscdevice.h>
#include <linux/interrupt.h>
#include <linux/ioport.h>
#include <linux/timer.h>
#include <linux/smp_lock.h>
#include <linux/io.h>
#include <asm/irq.h>
#include <asm/ebus.h>
#include <asm/oplib.h>
#include <asm/uaccess.h>
#include <asm/watchdog.h>
#define WD_OBPNAME "watchdog"
#define WD_BADMODEL "SUNW,501-5336"
#define WD_BTIMEOUT (jiffies + (HZ * 1000))
#define WD_BLIMIT 0xFFFF
#define WD0_DEVNAME "watchdog0"
#define WD1_DEVNAME "watchdog1"
#define WD2_DEVNAME "watchdog2"
#define WD0_MINOR 212
#define WD1_MINOR 213
#define WD2_MINOR 214
/* Internal driver definitions
*/
#define WD0_ID 0 /* Watchdog0 */
#define WD1_ID 1 /* Watchdog1 */
#define WD2_ID 2 /* Watchdog2 */
#define WD_NUMDEVS 3 /* Device contains 3 timers */
#define WD_INTR_OFF 0 /* Interrupt disable value */
#define WD_INTR_ON 1 /* Interrupt enable value */
#define WD_STAT_INIT 0x01 /* Watchdog timer is initialized */
#define WD_STAT_BSTOP 0x02 /* Watchdog timer is brokenstopped */
#define WD_STAT_SVCD 0x04 /* Watchdog interrupt occurred */
/* Register value definitions
*/
#define WD0_INTR_MASK 0x01 /* Watchdog device interrupt masks */
#define WD1_INTR_MASK 0x02
#define WD2_INTR_MASK 0x04
#define WD_S_RUNNING 0x01 /* Watchdog device status running */
#define WD_S_EXPIRED 0x02 /* Watchdog device status expired */
/* Sun uses Altera PLD EPF8820ATC144-4
* providing three hardware watchdogs:
*
* 1) RIC - sends an interrupt when triggered
* 2) XIR - asserts XIR_B_RESET when triggered, resets CPU
* 3) POR - asserts POR_B_RESET when triggered, resets CPU, backplane, board
*
*** Timer register block definition (struct wd_timer_regblk)
*
* dcntr and limit registers (halfword access):
* -------------------
* | 15 | ...| 1 | 0 |
* -------------------
* |- counter val -|
* -------------------
* dcntr - Current 16-bit downcounter value.
* When downcounter reaches '0' watchdog expires.
* Reading this register resets downcounter with 'limit' value.
* limit - 16-bit countdown value in 1/10th second increments.
* Writing this register begins countdown with input value.
* Reading from this register does not affect counter.
* NOTES: After watchdog reset, dcntr and limit contain '1'
*
* status register (byte access):
* ---------------------------
* | 7 | ... | 2 | 1 | 0 |
* --------------+------------
* |- UNUSED -| EXP | RUN |
* ---------------------------
* status- Bit 0 - Watchdog is running
* Bit 1 - Watchdog has expired
*
*** PLD register block definition (struct wd_pld_regblk)
*
* intr_mask register (byte access):
* ---------------------------------
* | 7 | ... | 3 | 2 | 1 | 0 |
* +-------------+------------------
* |- UNUSED -| WD3 | WD2 | WD1 |
* ---------------------------------
* WD3 - 1 == Interrupt disabled for watchdog 3
* WD2 - 1 == Interrupt disabled for watchdog 2
* WD1 - 1 == Interrupt disabled for watchdog 1
*
* pld_status register (byte access):
* UNKNOWN, MAGICAL MYSTERY REGISTER
*
*/
#define WD_TIMER_REGSZ 16
#define WD0_OFF 0
#define WD1_OFF (WD_TIMER_REGSZ * 1)
#define WD2_OFF (WD_TIMER_REGSZ * 2)
#define PLD_OFF (WD_TIMER_REGSZ * 3)
#define WD_DCNTR 0x00
#define WD_LIMIT 0x04
#define WD_STATUS 0x08
#define PLD_IMASK (PLD_OFF + 0x00)
#define PLD_STATUS (PLD_OFF + 0x04)
/* Individual timer structure
*/
struct wd_timer {
__u16 timeout;
__u8 intr_mask;
unsigned char runstatus;
void __iomem *regs;
};
/* Device structure
*/
struct wd_device {
int irq;
spinlock_t lock;
unsigned char isbaddoggie; /* defective PLD */
unsigned char opt_enable;
unsigned char opt_reboot;
unsigned short opt_timeout;
unsigned char initialized;
struct wd_timer watchdog[WD_NUMDEVS];
void __iomem *regs;
};
static struct wd_device wd_dev = {
0, __SPIN_LOCK_UNLOCKED(wd_dev.lock), 0, 0, 0, 0,
};
static struct timer_list wd_timer;
static int wd0_timeout = 0;
static int wd1_timeout = 0;
static int wd2_timeout = 0;
#ifdef MODULE
module_param (wd0_timeout, int, 0);
MODULE_PARM_DESC(wd0_timeout, "Default watchdog0 timeout in 1/10secs");
module_param (wd1_timeout, int, 0);
MODULE_PARM_DESC(wd1_timeout, "Default watchdog1 timeout in 1/10secs");
module_param (wd2_timeout, int, 0);
MODULE_PARM_DESC(wd2_timeout, "Default watchdog2 timeout in 1/10secs");
MODULE_AUTHOR
("Eric Brower <ebrower@usa.net>");
MODULE_DESCRIPTION
("Hardware watchdog driver for Sun Microsystems CP1400/1500");
MODULE_LICENSE("GPL");
MODULE_SUPPORTED_DEVICE
("watchdog");
#endif /* ifdef MODULE */
/* Forward declarations of internal methods
*/
#ifdef WD_DEBUG
static void wd_dumpregs(void);
#endif
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 13:55:46 +00:00
static irqreturn_t wd_interrupt(int irq, void *dev_id);
static void wd_toggleintr(struct wd_timer* pTimer, int enable);
static void wd_pingtimer(struct wd_timer* pTimer);
static void wd_starttimer(struct wd_timer* pTimer);
static void wd_resetbrokentimer(struct wd_timer* pTimer);
static void wd_stoptimer(struct wd_timer* pTimer);
static void wd_brokentimer(unsigned long data);
static int wd_getstatus(struct wd_timer* pTimer);
/* PLD expects words to be written in LSB format,
* so we must flip all words prior to writing them to regs
*/
static inline unsigned short flip_word(unsigned short word)
{
return ((word & 0xff) << 8) | ((word >> 8) & 0xff);
}
#define wd_writew(val, addr) (writew(flip_word(val), addr))
#define wd_readw(addr) (flip_word(readw(addr)))
#define wd_writeb(val, addr) (writeb(val, addr))
#define wd_readb(addr) (readb(addr))
/* CP1400s seem to have broken PLD implementations--
* the interrupt_mask register cannot be written, so
* no timer interrupts can be masked within the PLD.
*/
static inline int wd_isbroken(void)
{
/* we could test this by read/write/read/restore
* on the interrupt mask register only if OBP
* 'watchdog-enable?' == FALSE, but it seems
* ubiquitous on CP1400s
*/
char val[32];
prom_getproperty(prom_root_node, "model", val, sizeof(val));
return((!strcmp(val, WD_BADMODEL)) ? 1 : 0);
}
/* Retrieve watchdog-enable? option from OBP
* Returns 0 if false, 1 if true
*/
static inline int wd_opt_enable(void)
{
int opt_node;
opt_node = prom_getchild(prom_root_node);
opt_node = prom_searchsiblings(opt_node, "options");
return((-1 == prom_getint(opt_node, "watchdog-enable?")) ? 0 : 1);
}
/* Retrieve watchdog-reboot? option from OBP
* Returns 0 if false, 1 if true
*/
static inline int wd_opt_reboot(void)
{
int opt_node;
opt_node = prom_getchild(prom_root_node);
opt_node = prom_searchsiblings(opt_node, "options");
return((-1 == prom_getint(opt_node, "watchdog-reboot?")) ? 0 : 1);
}
/* Retrieve watchdog-timeout option from OBP
* Returns OBP value, or 0 if not located
*/
static inline int wd_opt_timeout(void)
{
int opt_node;
char value[32];
char *p = value;
opt_node = prom_getchild(prom_root_node);
opt_node = prom_searchsiblings(opt_node, "options");
opt_node = prom_getproperty(opt_node,
"watchdog-timeout",
value,
sizeof(value));
if(-1 != opt_node) {
/* atoi implementation */
for(opt_node = 0; /* nop */; p++) {
if(*p >= '0' && *p <= '9') {
opt_node = (10*opt_node)+(*p-'0');
}
else {
break;
}
}
}
return((-1 == opt_node) ? (0) : (opt_node));
}
static int wd_open(struct inode *inode, struct file *f)
{
switch(iminor(inode))
{
case WD0_MINOR:
f->private_data = &wd_dev.watchdog[WD0_ID];
break;
case WD1_MINOR:
f->private_data = &wd_dev.watchdog[WD1_ID];
break;
case WD2_MINOR:
f->private_data = &wd_dev.watchdog[WD2_ID];
break;
default:
return(-ENODEV);
}
/* Register IRQ on first open of device */
if(0 == wd_dev.initialized)
{
if (request_irq(wd_dev.irq,
&wd_interrupt,
IRQF_SHARED,
WD_OBPNAME,
(void *)wd_dev.regs)) {
printk("%s: Cannot register IRQ %d\n",
WD_OBPNAME, wd_dev.irq);
return(-EBUSY);
}
wd_dev.initialized = 1;
}
return(nonseekable_open(inode, f));
}
static int wd_release(struct inode *inode, struct file *file)
{
return 0;
}
static int wd_ioctl(struct inode *inode, struct file *file,
unsigned int cmd, unsigned long arg)
{
int setopt = 0;
struct wd_timer* pTimer = (struct wd_timer*)file->private_data;
void __user *argp = (void __user *)arg;
struct watchdog_info info = {
0,
0,
"Altera EPF8820ATC144-4"
};
if(NULL == pTimer) {
return(-EINVAL);
}
switch(cmd)
{
/* Generic Linux IOCTLs */
case WDIOC_GETSUPPORT:
if(copy_to_user(argp, &info, sizeof(struct watchdog_info))) {
return(-EFAULT);
}
break;
case WDIOC_GETSTATUS:
case WDIOC_GETBOOTSTATUS:
if (put_user(0, (int __user *)argp))
return -EFAULT;
break;
case WDIOC_KEEPALIVE:
wd_pingtimer(pTimer);
break;
case WDIOC_SETOPTIONS:
if(copy_from_user(&setopt, argp, sizeof(unsigned int))) {
return -EFAULT;
}
if(setopt & WDIOS_DISABLECARD) {
if(wd_dev.opt_enable) {
printk(
"%s: cannot disable watchdog in ENABLED mode\n",
WD_OBPNAME);
return(-EINVAL);
}
wd_stoptimer(pTimer);
}
else if(setopt & WDIOS_ENABLECARD) {
wd_starttimer(pTimer);
}
else {
return(-EINVAL);
}
break;
/* Solaris-compatible IOCTLs */
case WIOCGSTAT:
setopt = wd_getstatus(pTimer);
if(copy_to_user(argp, &setopt, sizeof(unsigned int))) {
return(-EFAULT);
}
break;
case WIOCSTART:
wd_starttimer(pTimer);
break;
case WIOCSTOP:
if(wd_dev.opt_enable) {
printk("%s: cannot disable watchdog in ENABLED mode\n",
WD_OBPNAME);
return(-EINVAL);
}
wd_stoptimer(pTimer);
break;
default:
return(-EINVAL);
}
return(0);
}
static long wd_compat_ioctl(struct file *file, unsigned int cmd,
unsigned long arg)
{
int rval = -ENOIOCTLCMD;
switch (cmd) {
/* solaris ioctls are specific to this driver */
case WIOCSTART:
case WIOCSTOP:
case WIOCGSTAT:
lock_kernel();
rval = wd_ioctl(file->f_path.dentry->d_inode, file, cmd, arg);
unlock_kernel();
break;
/* everything else is handled by the generic compat layer */
default:
break;
}
return rval;
}
static ssize_t wd_write(struct file *file,
const char __user *buf,
size_t count,
loff_t *ppos)
{
struct wd_timer* pTimer = (struct wd_timer*)file->private_data;
if(NULL == pTimer) {
return(-EINVAL);
}
if (count) {
wd_pingtimer(pTimer);
return 1;
}
return 0;
}
static ssize_t wd_read(struct file * file, char __user *buffer,
size_t count, loff_t *ppos)
{
#ifdef WD_DEBUG
wd_dumpregs();
return(0);
#else
return(-EINVAL);
#endif /* ifdef WD_DEBUG */
}
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 13:55:46 +00:00
static irqreturn_t wd_interrupt(int irq, void *dev_id)
{
/* Only WD0 will interrupt-- others are NMI and we won't
* see them here....
*/
spin_lock_irq(&wd_dev.lock);
if((unsigned long)wd_dev.regs == (unsigned long)dev_id)
{
wd_stoptimer(&wd_dev.watchdog[WD0_ID]);
wd_dev.watchdog[WD0_ID].runstatus |= WD_STAT_SVCD;
}
spin_unlock_irq(&wd_dev.lock);
return IRQ_HANDLED;
}
static const struct file_operations wd_fops = {
.owner = THIS_MODULE,
.ioctl = wd_ioctl,
.compat_ioctl = wd_compat_ioctl,
.open = wd_open,
.write = wd_write,
.read = wd_read,
.release = wd_release,
};
static struct miscdevice wd0_miscdev = { WD0_MINOR, WD0_DEVNAME, &wd_fops };
static struct miscdevice wd1_miscdev = { WD1_MINOR, WD1_DEVNAME, &wd_fops };
static struct miscdevice wd2_miscdev = { WD2_MINOR, WD2_DEVNAME, &wd_fops };
#ifdef WD_DEBUG
static void wd_dumpregs(void)
{
/* Reading from downcounters initiates watchdog countdown--
* Example is included below for illustration purposes.
*/
int i;
printk("%s: dumping register values\n", WD_OBPNAME);
for(i = WD0_ID; i < WD_NUMDEVS; ++i) {
/* printk("\t%s%i: dcntr at 0x%lx: 0x%x\n",
* WD_OBPNAME,
* i,
* (unsigned long)(&wd_dev.watchdog[i].regs->dcntr),
* readw(&wd_dev.watchdog[i].regs->dcntr));
*/
printk("\t%s%i: limit at 0x%lx: 0x%x\n",
WD_OBPNAME,
i,
(unsigned long)(&wd_dev.watchdog[i].regs->limit),
readw(&wd_dev.watchdog[i].regs->limit));
printk("\t%s%i: status at 0x%lx: 0x%x\n",
WD_OBPNAME,
i,
(unsigned long)(&wd_dev.watchdog[i].regs->status),
readb(&wd_dev.watchdog[i].regs->status));
printk("\t%s%i: driver status: 0x%x\n",
WD_OBPNAME,
i,
wd_getstatus(&wd_dev.watchdog[i]));
}
printk("\tintr_mask at %p: 0x%x\n",
wd_dev.regs + PLD_IMASK,
readb(wd_dev.regs + PLD_IMASK));
printk("\tpld_status at %p: 0x%x\n",
wd_dev.regs + PLD_STATUS,
readb(wd_dev.regs + PLD_STATUS));
}
#endif
/* Enable or disable watchdog interrupts
* Because of the CP1400 defect this should only be
* called during initialzation or by wd_[start|stop]timer()
*
* pTimer - pointer to timer device, or NULL to indicate all timers
* enable - non-zero to enable interrupts, zero to disable
*/
static void wd_toggleintr(struct wd_timer* pTimer, int enable)
{
unsigned char curregs = wd_readb(wd_dev.regs + PLD_IMASK);
unsigned char setregs =
(NULL == pTimer) ?
(WD0_INTR_MASK | WD1_INTR_MASK | WD2_INTR_MASK) :
(pTimer->intr_mask);
(WD_INTR_ON == enable) ?
(curregs &= ~setregs):
(curregs |= setregs);
wd_writeb(curregs, wd_dev.regs + PLD_IMASK);
return;
}
/* Reset countdown timer with 'limit' value and continue countdown.
* This will not start a stopped timer.
*
* pTimer - pointer to timer device
*/
static void wd_pingtimer(struct wd_timer* pTimer)
{
if (wd_readb(pTimer->regs + WD_STATUS) & WD_S_RUNNING) {
wd_readw(pTimer->regs + WD_DCNTR);
}
}
/* Stop a running watchdog timer-- the timer actually keeps
* running, but the interrupt is masked so that no action is
* taken upon expiration.
*
* pTimer - pointer to timer device
*/
static void wd_stoptimer(struct wd_timer* pTimer)
{
if(wd_readb(pTimer->regs + WD_STATUS) & WD_S_RUNNING) {
wd_toggleintr(pTimer, WD_INTR_OFF);
if(wd_dev.isbaddoggie) {
pTimer->runstatus |= WD_STAT_BSTOP;
wd_brokentimer((unsigned long)&wd_dev);
}
}
}
/* Start a watchdog timer with the specified limit value
* If the watchdog is running, it will be restarted with
* the provided limit value.
*
* This function will enable interrupts on the specified
* watchdog.
*
* pTimer - pointer to timer device
* limit - limit (countdown) value in 1/10th seconds
*/
static void wd_starttimer(struct wd_timer* pTimer)
{
if(wd_dev.isbaddoggie) {
pTimer->runstatus &= ~WD_STAT_BSTOP;
}
pTimer->runstatus &= ~WD_STAT_SVCD;
wd_writew(pTimer->timeout, pTimer->regs + WD_LIMIT);
wd_toggleintr(pTimer, WD_INTR_ON);
}
/* Restarts timer with maximum limit value and
* does not unset 'brokenstop' value.
*/
static void wd_resetbrokentimer(struct wd_timer* pTimer)
{
wd_toggleintr(pTimer, WD_INTR_ON);
wd_writew(WD_BLIMIT, pTimer->regs + WD_LIMIT);
}
/* Timer device initialization helper.
* Returns 0 on success, other on failure
*/
static int wd_inittimer(int whichdog)
{
struct miscdevice *whichmisc;
void __iomem *whichregs;
char whichident[8];
int whichmask;
__u16 whichlimit;
switch(whichdog)
{
case WD0_ID:
whichmisc = &wd0_miscdev;
strcpy(whichident, "RIC");
whichregs = wd_dev.regs + WD0_OFF;
whichmask = WD0_INTR_MASK;
whichlimit= (0 == wd0_timeout) ?
(wd_dev.opt_timeout):
(wd0_timeout);
break;
case WD1_ID:
whichmisc = &wd1_miscdev;
strcpy(whichident, "XIR");
whichregs = wd_dev.regs + WD1_OFF;
whichmask = WD1_INTR_MASK;
whichlimit= (0 == wd1_timeout) ?
(wd_dev.opt_timeout):
(wd1_timeout);
break;
case WD2_ID:
whichmisc = &wd2_miscdev;
strcpy(whichident, "POR");
whichregs = wd_dev.regs + WD2_OFF;
whichmask = WD2_INTR_MASK;
whichlimit= (0 == wd2_timeout) ?
(wd_dev.opt_timeout):
(wd2_timeout);
break;
default:
printk("%s: %s: invalid watchdog id: %i\n",
WD_OBPNAME, __func__, whichdog);
return(1);
}
if(0 != misc_register(whichmisc))
{
return(1);
}
wd_dev.watchdog[whichdog].regs = whichregs;
wd_dev.watchdog[whichdog].timeout = whichlimit;
wd_dev.watchdog[whichdog].intr_mask = whichmask;
wd_dev.watchdog[whichdog].runstatus &= ~WD_STAT_BSTOP;
wd_dev.watchdog[whichdog].runstatus |= WD_STAT_INIT;
printk("%s%i: %s hardware watchdog [%01i.%i sec] %s\n",
WD_OBPNAME,
whichdog,
whichident,
wd_dev.watchdog[whichdog].timeout / 10,
wd_dev.watchdog[whichdog].timeout % 10,
(0 != wd_dev.opt_enable) ? "in ENABLED mode" : "");
return(0);
}
/* Timer method called to reset stopped watchdogs--
* because of the PLD bug on CP1400, we cannot mask
* interrupts within the PLD so me must continually
* reset the timers ad infinitum.
*/
static void wd_brokentimer(unsigned long data)
{
struct wd_device* pDev = (struct wd_device*)data;
int id, tripped = 0;
/* kill a running timer instance, in case we
* were called directly instead of by kernel timer
*/
if(timer_pending(&wd_timer)) {
del_timer(&wd_timer);
}
for(id = WD0_ID; id < WD_NUMDEVS; ++id) {
if(pDev->watchdog[id].runstatus & WD_STAT_BSTOP) {
++tripped;
wd_resetbrokentimer(&pDev->watchdog[id]);
}
}
if(tripped) {
/* there is at least one timer brokenstopped-- reschedule */
init_timer(&wd_timer);
wd_timer.expires = WD_BTIMEOUT;
add_timer(&wd_timer);
}
}
static int wd_getstatus(struct wd_timer* pTimer)
{
unsigned char stat = wd_readb(pTimer->regs + WD_STATUS);
unsigned char intr = wd_readb(wd_dev.regs + PLD_IMASK);
unsigned char ret = WD_STOPPED;
/* determine STOPPED */
if(0 == stat ) {
return(ret);
}
/* determine EXPIRED vs FREERUN vs RUNNING */
else if(WD_S_EXPIRED & stat) {
ret = WD_EXPIRED;
}
else if(WD_S_RUNNING & stat) {
if(intr & pTimer->intr_mask) {
ret = WD_FREERUN;
}
else {
/* Fudge WD_EXPIRED status for defective CP1400--
* IF timer is running
* AND brokenstop is set
* AND an interrupt has been serviced
* we are WD_EXPIRED.
*
* IF timer is running
* AND brokenstop is set
* AND no interrupt has been serviced
* we are WD_FREERUN.
*/
if(wd_dev.isbaddoggie && (pTimer->runstatus & WD_STAT_BSTOP)) {
if(pTimer->runstatus & WD_STAT_SVCD) {
ret = WD_EXPIRED;
}
else {
/* we could as well pretend we are expired */
ret = WD_FREERUN;
}
}
else {
ret = WD_RUNNING;
}
}
}
/* determine SERVICED */
if(pTimer->runstatus & WD_STAT_SVCD) {
ret |= WD_SERVICED;
}
return(ret);
}
static int __init wd_init(void)
{
int id;
struct linux_ebus *ebus = NULL;
struct linux_ebus_device *edev = NULL;
for_each_ebus(ebus) {
for_each_ebusdev(edev, ebus) {
if (!strcmp(edev->ofdev.node->name, WD_OBPNAME))
goto ebus_done;
}
}
ebus_done:
if(!edev) {
printk("%s: unable to locate device\n", WD_OBPNAME);
return -ENODEV;
}
wd_dev.regs =
ioremap(edev->resource[0].start, 4 * WD_TIMER_REGSZ); /* ? */
if(NULL == wd_dev.regs) {
printk("%s: unable to map registers\n", WD_OBPNAME);
return(-ENODEV);
}
/* initialize device structure from OBP parameters */
wd_dev.irq = edev->irqs[0];
wd_dev.opt_enable = wd_opt_enable();
wd_dev.opt_reboot = wd_opt_reboot();
wd_dev.opt_timeout = wd_opt_timeout();
wd_dev.isbaddoggie = wd_isbroken();
/* disable all interrupts unless watchdog-enabled? == true */
if(! wd_dev.opt_enable) {
wd_toggleintr(NULL, WD_INTR_OFF);
}
/* register miscellaneous devices */
for(id = WD0_ID; id < WD_NUMDEVS; ++id) {
if(0 != wd_inittimer(id)) {
printk("%s%i: unable to initialize\n", WD_OBPNAME, id);
}
}
/* warn about possible defective PLD */
if(wd_dev.isbaddoggie) {
init_timer(&wd_timer);
wd_timer.function = wd_brokentimer;
wd_timer.data = (unsigned long)&wd_dev;
wd_timer.expires = WD_BTIMEOUT;
printk("%s: PLD defect workaround enabled for model %s\n",
WD_OBPNAME, WD_BADMODEL);
}
return(0);
}
static void __exit wd_cleanup(void)
{
int id;
/* if 'watchdog-enable?' == TRUE, timers are not stopped
* when module is unloaded. All brokenstopped timers will
* also now eventually trip.
*/
for(id = WD0_ID; id < WD_NUMDEVS; ++id) {
if(WD_S_RUNNING == wd_readb(wd_dev.watchdog[id].regs + WD_STATUS)) {
if(wd_dev.opt_enable) {
printk(KERN_WARNING "%s%i: timer not stopped at release\n",
WD_OBPNAME, id);
}
else {
wd_stoptimer(&wd_dev.watchdog[id]);
if(wd_dev.watchdog[id].runstatus & WD_STAT_BSTOP) {
wd_resetbrokentimer(&wd_dev.watchdog[id]);
printk(KERN_WARNING
"%s%i: defect workaround disabled at release, "\
"timer expires in ~%01i sec\n",
WD_OBPNAME, id,
wd_readw(wd_dev.watchdog[id].regs + WD_LIMIT) / 10);
}
}
}
}
if(wd_dev.isbaddoggie && timer_pending(&wd_timer)) {
del_timer(&wd_timer);
}
if(0 != (wd_dev.watchdog[WD0_ID].runstatus & WD_STAT_INIT)) {
misc_deregister(&wd0_miscdev);
}
if(0 != (wd_dev.watchdog[WD1_ID].runstatus & WD_STAT_INIT)) {
misc_deregister(&wd1_miscdev);
}
if(0 != (wd_dev.watchdog[WD2_ID].runstatus & WD_STAT_INIT)) {
misc_deregister(&wd2_miscdev);
}
if(0 != wd_dev.initialized) {
free_irq(wd_dev.irq, (void *)wd_dev.regs);
}
iounmap(wd_dev.regs);
}
module_init(wd_init);
module_exit(wd_cleanup);