d/freebsd-skq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Apply a patch that has been lingering in my inbox for far too long:

Browse Source 
On a soekris Net5501, if you do a watchdog -t 16, followed by a watchdog
-t 0 to disable the watchdog, and then after some time (16s) re-enable
the watchdog the box reboots immediatly. This prevents also to stop and
restart watchdogd(8).

This is because when you stop the watchdog, the timer is not stoped,
only the hard reset is disabled. So when the timer has elapsed, the C2
event of the timer is set.

But when the hard reset is re-enabled, the event is not cleared and the
box reboots.

The attached patch stops and resets the counter when the watchdog is
disabled and do not disable the hard reset of the timer (if the timer
has elapsed it's too late).

Submitted by:	 Patrick Lamaizière
This commit is contained in:
Poul-Henning Kamp 2010-05-15 10:31:11 +00:00
parent c6bfb71a6f
commit 965df046e5
1 changed files with 13 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
sys/i386/i386/geode.c
View File

@ -208,14 +208,11 @@ geode_watchdog(void *foo __unused, u_int cmd, int *error)
static void
cs5536_watchdog(void *foo __unused, u_int cmd, int *error)
{
u_int u, p;
u_int u, p, s;
uint16_t a;
uint32_t m;
a = rdmsr(0x5140000d);
m = rdmsr(0x51400029);
m &= ~(1 << 24);
wrmsr(0x51400029, m);
u = cmd & WD_INTERVAL;
if (u >= 30 && u <= 44) {
@ -228,12 +225,24 @@ cs5536_watchdog(void *foo __unused, u_int cmd, int *error)
/* reset counter */
outw(a + 4, 0);
/* Arm reset mechanism */
m = rdmsr(0x51400029);
m |= (1 << 24);
wrmsr(0x51400029, m);
/* Start counter */
outw(a + 6, 0x8000);
*error = 0;
} else {
/*
* MFGPT_SETUP is write-once
* Check if the counter has been setup
*/
s = inw(a + 6);
if (s & (1 << 12)) {
/* Stop and reset counter */
outw(a + 6, 0);
outw(a + 4, 0);
}
}
}