dtrace discuss - Jun 2008 - Arg1 doesn''t hold return value in pid provider?

I am trying to trace the xti functions in a process. I am using the 
following script:

#!/usr/sbin/dtrace -s
pid$target::t_*:entry
{
}
pid$target::t_*:return
{
         printf("%x %d",arg0,arg1);
}


My problem is that the value of arg1 that I get is mostly 50 or 48, 
although I have seen large numbers like 642832 and 768712. The thing 
is, the t_* functions almost all return either a 0 or -1. I thought I 
must be missing something simple, like a copying or 32/64 bit thing, 
but I just don''t see what the problem is. Any ides?
-- 
blu

There are two rules in life:
Rule 1- Don''t tell people everything you know
----------------------------------------------------------------------
Brian Utterback - Solaris RPE, Sun Microsystems, Inc.
Ph:877-259-7345, Em:brian.utterback-at-ess-you-enn-dot-kom

Hi Brian,

Functions that end in a tail-call to another function can report an  
incorrect
return value in the arg1 variable. Take a look at the disassembly at the
location where you''re seeing the apparently invalid return values to  
confirm
that it''s a tail-call.

It can be a little confusing -- sorry about that.

Adam

On Jun 25, 2008, at 5:09 AM, Brian Utterback wrote:
> I am trying to trace the xti functions in a process. I am using the
> following script:
>
> #!/usr/sbin/dtrace -s
> pid$target::t_*:entry
> {
> }
> pid$target::t_*:return
> {
>         printf("%x %d",arg0,arg1);
> }
>
>
> My problem is that the value of arg1 that I get is mostly 50 or 48,
> although I have seen large numbers like 642832 and 768712. The thing
> is, the t_* functions almost all return either a 0 or -1. I thought I
> must be missing something simple, like a copying or 32/64 bit thing,
> but I just don''t see what the problem is. Any ides?
> -- 
> blu
>
> There are two rules in life:
> Rule 1- Don''t tell people everything you know
> ----------------------------------------------------------------------
> 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, Fishworks                        http://blogs.sun.com/ahl

That''s what it was. All of the t_* functions are just wrapper 
functions that call the corresponding _tx_* function. Darn tail calls.

Adam Leventhal wrote:
> Hi Brian,
> 
> Functions that end in a tail-call to another function can report an  
> incorrect
> return value in the arg1 variable. Take a look at the disassembly at the
> location where you''re seeing the apparently invalid return values
to
> confirm
> that it''s a tail-call.
> 
> It can be a little confusing -- sorry about that.
> 
> Adam
> 
> On Jun 25, 2008, at 5:09 AM, Brian Utterback wrote:
> 
>> I am trying to trace the xti functions in a process. I am using the
>> following script:
>>
>> #!/usr/sbin/dtrace -s
>> pid$target::t_*:entry
>> {
>> }
>> pid$target::t_*:return
>> {
>>         printf("%x %d",arg0,arg1);
>> }
>>
>>
>> My problem is that the value of arg1 that I get is mostly 50 or 48,
>> although I have seen large numbers like 642832 and 768712. The thing
>> is, the t_* functions almost all return either a 0 or -1. I thought I
>> must be missing something simple, like a copying or 32/64 bit thing,
>> but I just don''t see what the problem is. Any ides?
>> -- 
>> blu
>>
>> There are two rules in life:
>> Rule 1- Don''t tell people everything you know
>> ----------------------------------------------------------------------
>> 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, Fishworks                        http://blogs.sun.com/ahl
> 
> _______________________________________________
> dtrace-discuss mailing list
> dtrace-discuss at opensolaris.org
-- 
blu

There are two rules in life:
Rule 1- Don''t tell people everything you know
----------------------------------------------------------------------
Brian Utterback - Solaris RPE, Sun Microsystems, Inc.
Ph:877-259-7345, Em:brian.utterback-at-ess-you-enn-dot-kom

> Functions that end in a tail-call to another function can report an
> incorrect return value in the arg1 variable. Take a look at the
> disassembly at the location where you''re seeing the apparently
invalid
> return values to  confirm that it''s a tail-call.
Aaah, that exactly happened to me half a year ago, on the same t_*
functions. And I forgot it in the meantime. Would it be sensible to have
dtrace which would warn you when it detects tail-call optimized function
?

-- 
	Vlad
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 193 bytes
Desc: not available
URL:
<http://mail.opensolaris.org/pipermail/dtrace-discuss/attachments/20080626/b3da97fb/attachment.bin>

>> Functions that end in a tail-call to another function can report an
>> incorrect return value in the arg1 variable. Take a look at the
>> disassembly at the location where you''re seeing the apparently
>> invalid
>> return values to  confirm that it''s a tail-call.
>
> Aaah, that exactly happened to me half a year ago, on the same t_*
> functions. And I forgot it in the meantime. Would it be sensible to  
> have
> dtrace which would warn you when it detects tail-call optimized  
> function
> ?

It''s certainly the case that it''s never the right thing to
trace arg1
from a tail-call return probe. It would be tricky, but perhaps DTrace
could do something to warn against this. I suggest you file a bug.

Adam

--
Adam Leventhal, Fishworks                        http://blogs.sun.com/ahl

>> Aaah, that exactly happened to me half a year ago, on the same t_*
>> functions. And I forgot it in the meantime. Would it be sensible to
>> have dtrace which would warn you when it detects tail-call optimized
>> function ?
>
> It''s certainly the case that it''s never the right thing
to trace arg1
> from a tail-call return probe. It would be tricky, but perhaps DTrace
> could do something to warn against this. I suggest you file a bug.
6720127 dtrace could warn user that he''s tracing tail-call optimized
function

Thank you

-- 
	Vlad
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 193 bytes
Desc: not available
URL:
<http://mail.opensolaris.org/pipermail/dtrace-discuss/attachments/20080627/ed387f4a/attachment.bin>