calcru: runtime went backwards, RELENG_6, SMP

[Dmitry Morozovsky]

On Sat, 9 Jun 2007, Matthew Dillon wrote:

MD> :Hmm, i'm not sure I understand you right: what do you mean by 'kgdb live 
MD> :kernel'? I send break over serial console, and in ddb got
MD> :
MD> :db> print cpu_ticks
MD> :Symbol not found
MD> :
MD> :Sincerely,
MD> :D.Marck                                     [DM5020, MCK-RIPE, DM3-RIPN]
MD> 
MD>     I think it works the same on FreeBSD, so it would be something like:
MD> 
MD>     kgdb /kernel /dev/mem
MD> 
MD>     ^^^ NOTE! Dangerous! ^^^
MD> 
MD>     But I looked at the cvs logs and the variable didn't exist in FreeBSD-6,
MD>     so it wouldn't have helped anyway.
MD> 
MD>     It looks like it is using binuptime() in 6.x, and it also looks like
MD>     the tick calculations, e.g. rux_uticks, is based on the stat clock
MD>     interrupt, whereas the runtime calculation is using binuptime.  There
MD>     is no way those two could possibly be synchronized.  No chance
MD>     whatsoever.  Your only solution may be to upgrade to FreeBSD-7 which
MD>     uses an entirely different mechanism for the calculation (though one
MD>     that also seems flawed in its own way).
MD> 
MD>     Alternatively you could just remove the error message from the kernel
MD>     entirely and not worry about it.  It's a printf around line 774
MD>     in /usr/src/sys/kern/kern_resource.c (in FreeBSD-6.x).

Well, I can of course shut the kernel up, but kernel time stability is still my 
concern. I run ntpd there and while sometimes it seems stable (well, sorta: 
drift are within several seconds...) there are cases of half-a-minute time 
steps. 
 

Sincerely,
D.Marck                                     [DM5020, MCK-RIPE, DM3-RIPN]
------------------------------------------------------------------------
*** Dmitry Morozovsky --- D.Marck --- Wild Woozle --- marck at rinet.ru ***
------------------------------------------------------------------------