This patch has been verified to apply cleanly to 2.5.30
Bitops are used with on the timer_alive variable. Therefore, timer_alive
needs to be "unsigned long" not "int".
drivers/char/wdt977.c | 2 +-
1 files changed, 1 insertion, 1 deletion
diff -urN orig/drivers/char/wdt977.c linux/drivers/char/wdt977.c
--- orig/drivers/char/wdt977.c Sat May 25 23:13:25 2002
+++ linux/drivers/char/wdt977.c Wed Jun 12 14:13:47 2002
@@ -39,7 +39,7 @@
static int timeout = DEFAULT_TIMEOUT*60; /* TO in seconds from user */
static int timeoutM = DEFAULT_TIMEOUT; /* timeout in minutes */
-static int timer_alive;
+static unsigned long timer_alive;
static int testmode;
MODULE_PARM(timeout, "i");
On Sat, Aug 03, 2002 at 07:16:18PM +0100, Russell King wrote:
> diff -urN orig/drivers/char/wdt977.c linux/drivers/char/wdt977.c
> --- orig/drivers/char/wdt977.c Sat May 25 23:13:25 2002
> +++ linux/drivers/char/wdt977.c Wed Jun 12 14:13:47 2002
> @@ -39,7 +39,7 @@
>
> static int timeout = DEFAULT_TIMEOUT*60; /* TO in seconds from user */
> static int timeoutM = DEFAULT_TIMEOUT; /* timeout in minutes */
> -static int timer_alive;
> +static unsigned long timer_alive;
> static int testmode;
>
> MODULE_PARM(timeout, "i");
Not sure how much people will care, but this (and other funtastic) changes
are in my set of watchdog patches that I'll be making a big push to get
applied soon. I haven't updated them to 2.4.19 or 2.5.30 yet but I'm hoping
to soon. When I do, they'll be up at http://osinvestor.com/wd/ .
Regards,
Rob Radez
On Sat, Aug 03, 2002 at 07:16:18PM +0100, Russell King wrote:
> Bitops are used with on the timer_alive variable. Therefore, timer_alive
> needs to be "unsigned long" not "int".
How about:
> -static int timer_alive;
> +static bitmap_member(timer_alive, 1);
In this way it is obviously used for bitops.
Rusty's more informative DECLARE_BITMAP (IIRC) patch has not yet been
accepted by Linus, otherwise it would better tell people that this is
actually a macro without the need to parse the declaration first.
Sam