Re: `top` reports all CPU states as 0% on i386

Jason George wrote:

>>This box is a fairly fresh install of 5.2-RC2, running on Pentium
>>200MMX. The line from top reads as follows:
>>
>>CPU states:  0.0% user,  0.0% nice,  0.0% system,  0.0% interrupt,  0.0% idle
>>
>>I first noticed it when my cacti/rrdtool snmp graphs for CPU usage
>>showed 0 across the board.
>>
>>This afternoon I cvsup'd it to -CURRENT and rebuild kernel and world.
>>After rebooting the problem remained.
>>
>>Is this a known problem?
>>
>>    
>>
>
>I asked the same question Dec 11 as I encountered the same issue.
>A number of people responded in private email, suggesting that the kernel 
>was not synchronized properly with the userland utilities.  I doubted that, 
>as I always rebuild and install world when I rebuild a kernel.
>
>In the last month, I have cvsup'ed and rebuilt both the kernel + world at 
>least a dozen times since that time.  The problem still exists.
>
>My machine is a Dual Pentium Pro with a basically-stock SMP kernel, except 
>that I have ULE enabled. 
>
>--J 
>_______________________________________________
>freebsd-current_at_freebsd.org mailing list
>http://lists.freebsd.org/mailman/listinfo/freebsd-current
>To unsubscribe, send any mail to "freebsd-current-unsubscribe_at_freebsd.org"
>
>  
>
I've had the same exact problem several times in the past...pinpointing 
it to somewhere between post 5.1-RELEASE and 5.2-BETA as where it 
started.  5.1-RELEASE and subsequent 5.1-CURRENT through mid-November 
didn't seem to have this issue (for me of course).  System is an IBM 
Netfinity 4500R- 2x PIII SMP, ServeRAID 3L, 1G RAM.  If I'm not 
mistaken, the other instances of people experiencing this were also SMP 
boxes.

I'm not sure at what point it 'resolved itself,' but it's possible my 
kernel and world were out of sync, or that I skipped doing a 
mergemaster.... is now working with 5.2-CURRENT as of Jan 1....

Not sure this helps in any immediate fashion, but another data point....

Scott