On Sun, Jan 21, 2001 at 08:44:33PM -0600, Bob Nestor wrote:
> John Valdes wrote:
> 
> >Does NetBSD/macppc have any way to compensate for the fact the Mac
> >(normally) stores local time in the hardware clock?
> 
> Other ports have a similar problem.  Setting RTC_OFFSET would solve the 
> problem, but it seems that it's not writeable via sysctl.

Hmmm, I tried setting "options RTC_OFFSET=360" in the kernel config
file, but it didn't seem to have any effect.  Well, "sysctl
kern.rtc_offset" now reports 360 rather than 0, but the system clock
was unaffected.  A quick grep of /usr/src/arch/macppc/macppc/clock.c
didn't find any reference to rtc_offset, so it looks like this option
is ignored in the macppc port.  This was easy to patch though (see
below).  Actually, checking now, I see there are a couple of PR's on
this already...

> On the i386 
> port I got around this by using the instructions in the INSTALL document 
> for changing the address for some PCI devices.  Basically you boot up the 
> system and get it to drop into the debugger. Then you modify the 
> RTC_OFFSET value and write it back out.  Thereafter the NetBSD system 
> applies the RTC_OFFSET to the hardware clock and all is well -- until 
> Daylight Savings Time. ;)

I think I'll just compile two kernels; one for Standard Time, and one
for DST... :)

John

--------------------------------------------------------------------

--- /usr/src/sys/arch/macppc/macppc/clock.c.FCS	Wed Nov  1 10:01:50 2000
+++ /usr/src/sys/arch/macppc/macppc/clock.c	Fri Feb  2 13:02:17 2001
@@ -80,7 +80,7 @@
 		return;
 	}
 	clockinitted = 1;
-	time.tv_sec = rtc_time - DIFF19041970;
+	time.tv_sec = rtc_time - DIFF19041970 + rtc_offset * 60;
 
 	deltat = time.tv_sec - base;
 	if (deltat < 0)
@@ -104,7 +104,7 @@
 	u_int rtc_time;
 
 	if (clockinitted) {
-		rtc_time = time.tv_sec + DIFF19041970;
+		rtc_time = time.tv_sec + DIFF19041970 - rtc_offset * 60;
 		adb_set_date_time(rtc_time);
 	}
 #endif