dtrace discuss - Feb 2007 - How to print kernel global variable.

I am doing some testing on the adjtime syscall, and I want to
dump out the value of the kernel variable "timedelta", and I
am failing at it. Some of my attempts got "Module is no longer
loaded" and some got "Symbol table entry is not a data object"

I found some info that dtrace can print out kernel global
variable, so what am I missing here?
-- 
blu

"Remember ''A Thousand Points of Light''? With a network,
we now have
a thousand points of failure."
----------------------------------------------------------------------
Brian Utterback - Solaris RPE, Sun Microsystems, Inc.
Ph:877-259-7345, Em:brian.utterback-at-ess-you-enn-dot-kom

Hi Brian,

The problem is that DTrace won''t let you use a kernel variable if it
can''t
infer the type. Since ''timedelta'' is defined in an assembly
file, we don''t
have CTF data. Feel free to file a bug -- this should be made to work.

I was able to trace ''timedelta'' by getting its address using
mdb and plugging
that into my D script:

# echo timedelta=K | mdb -k
                fffffffffbc02930 
# dtrace -n BEGIN''{ trace(*(int64_t *)0xfffffffffbc02930ULL);
}''
dtrace: description ''BEGIN'' matched 1 probe
CPU     ID                    FUNCTION:NAME
  1      1                           :BEGIN                 0

Adam

On Tue, Feb 13, 2007 at 03:05:03PM -0500, Brian Utterback
wrote:
> I am doing some testing on the adjtime syscall, and I want to
> dump out the value of the kernel variable "timedelta", and I
> am failing at it. Some of my attempts got "Module is no longer
> loaded" and some got "Symbol table entry is not a data
object"
> 
> I found some info that dtrace can print out kernel global
> variable, so what am I missing here?
> -- 
> blu
> 
> "Remember ''A Thousand Points of Light''? With a
network, we now have
> a thousand points of failure."
> ----------------------------------------------------------------------
> Brian Utterback - Solaris RPE, Sun Microsystems, Inc.
> Ph:877-259-7345, Em:brian.utterback-at-ess-you-enn-dot-kom
> _______________________________________________
> dtrace-discuss mailing list
> dtrace-discuss at opensolaris.org
-- 
Adam Leventhal, Solaris Kernel Development       http://blogs.sun.com/ahl

Hmm. IS there anyway to tell dtrace what the type is? Something like the 
''-t'' flag to print in mdb?
Maybe ''int64_t timedelta'' perhaps?

Adam Leventhal wrote:
> Hi Brian,
>
> The problem is that DTrace won''t let you use a kernel variable if
it can''t
> infer the type. Since ''timedelta'' is defined in an
assembly file, we don''t
> have CTF data. Feel free to file a bug -- this should be made to work.
>
> I was able to trace ''timedelta'' by getting its address
using mdb and plugging
> that into my D script:
>
> # echo timedelta=K | mdb -k
>                 fffffffffbc02930 
> # dtrace -n BEGIN''{ trace(*(int64_t *)0xfffffffffbc02930ULL);
}''
> dtrace: description ''BEGIN'' matched 1 probe
> CPU     ID                    FUNCTION:NAME
>   1      1                           :BEGIN                 0
>
> Adam
>
> On Tue, Feb 13, 2007 at 03:05:03PM -0500, Brian Utterback wrote:
>   
>> I am doing some testing on the adjtime syscall, and I want to
>> dump out the value of the kernel variable "timedelta", and I
>> am failing at it. Some of my attempts got "Module is no longer
>> loaded" and some got "Symbol table entry is not a data
object"
>>
>> I found some info that dtrace can print out kernel global
>> variable, so what am I missing here?
>> -- 
>> blu
>>
>> "Remember ''A Thousand Points of Light''? With a
network, we now have
>> a thousand points of failure."
>> ----------------------------------------------------------------------
>> Brian Utterback - Solaris RPE, Sun Microsystems, Inc.
>> Ph:877-259-7345, Em:brian.utterback-at-ess-you-enn-dot-kom
>> _______________________________________________
>> dtrace-discuss mailing list
>> dtrace-discuss at opensolaris.org
>>     
>
>