Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 28 Jul 1996 04:52:50 +0800
From:      Peter Wemm <peter@spinner.dialix.com>
To:        freebsd-smp@freefall.freebsd.org
Subject:   Re: cvs commit: sys/i386/i386 microtime.s 
Message-ID:  <199607272052.EAA29937@spinner.DIALix.COM>
In-Reply-To: Your message of "Sat, 27 Jul 1996 13:39:43 MST." <199607272039.NAA01667@freefall.freebsd.org> 

next in thread | previous in thread | raw e-mail | index | archive | help
Peter Wemm wrote:
> peter       96/07/27 13:39:41
> 
>   Modified:    i386/i386  microtime.s
>   Log:
>   make this compile again in the non-SMP case.

BTW, has anybody got any bright ideas about the pentium_microtime code 
under SMP?  I've not looked to closely, but it seems the reason this has 
been disabled is because of individual calibration and offset issues.  And 
if it's going to be disabled here, shouldn't it also be disabled in 
clock.[ch] as well? Otherwise things like the counter adjustments are 
going to screw up when they update one pentium's clock but not the other?  
Perhaps some of those negative time counts on the console are caused by 
this.. How many people getting these are also running xntpd? (and hence, 
adjtime())

Also, wasn't the un-official word from the IAL something to the effect of 
that they were distressed that we were using the pentium counter for 
somthing that it was never meant to be?

Cheers,
-Peter





Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199607272052.EAA29937>