dtrace discuss - Jul 2011 - isenabled probles broken on Mac OS X Lion

G''Day,

The way isenabled probes are compiled on Mac OS X Lion has changed in a way
that dosen''t compile (at least, with some compilers).  Here''s
the test to
try:

$ cat -n liontest.d
     1 provider lion {
     2 probe hello(const char *world);
     3 };
$ dtrace -h -s liontest.d -o liontest.h

On Snow Leopard this output:

$ cat -n liontest.h
[...]
    24 #define LION_HELLO_ENABLED() \
    25 __dtrace_isenabled$lion$hello$v1()

On Lion that is now:

$ can -n liontest.h
[...]
    26 #define LION_HELLO_ENABLED() \
    27        ({ int _r = __dtrace_isenabled$lion$hello$v1(); \
    28                __asm__ volatile(""); \
    29                _r; })

Which doesn''t compile when used in if statements (the use case for
isenabled).  Which means, some software that has DTrace probes is now not
compiling on Lion.

Another difference is the addition of:

#if !defined(DTRACE_PROBES_DISABLED) || !DTRACE_PROBES_DISABLED

but I don''t think that is breaking anything.

I''ve attached both header files from the OS versions.  Does Apple know
of
this?

thanks,

Brendan

-- 
Brendan Gregg, Joyent                      http://dtrace.org/blogs/brendan
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.opensolaris.org/pipermail/dtrace-discuss/attachments/20110727/62ff632b/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: lion.h
Type: text/x-chdr
Size: 1061 bytes
Desc: not available
URL:
<http://mail.opensolaris.org/pipermail/dtrace-discuss/attachments/20110727/62ff632b/attachment.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: snowleopard.h
Type: text/x-chdr
Size: 711 bytes
Desc: not available
URL:
<http://mail.opensolaris.org/pipermail/dtrace-discuss/attachments/20110727/62ff632b/attachment-0001.bin>

I''m not able to reproduce, what compiler are you using?

	James M

On Jul 27, 2011, at 4:00 PM, Brendan Gregg wrote:
> G''Day,
> 
> The way isenabled probes are compiled on Mac OS X Lion has changed in a way
that dosen''t compile (at least, with some compilers).  Here''s
the test to try:
> 
> $ cat -n liontest.d 
>      1	provider lion {
>      2		probe hello(const char *world);
>      3	};
> $ dtrace -h -s liontest.d -o liontest.h
> 
> On Snow Leopard this output:
> 
> $ cat -n liontest.h
> [...]
>     24	#define	LION_HELLO_ENABLED() \
>     25		__dtrace_isenabled$lion$hello$v1()
> 
> On Lion that is now:
> 
> $ can -n liontest.h
> [...]
>     26	#define LION_HELLO_ENABLED() \
>     27	        ({ int _r = __dtrace_isenabled$lion$hello$v1(); \
>     28	                __asm__ volatile(""); \
>     29	                _r; })
> 
> Which doesn''t compile when used in if statements (the use case for
isenabled).  Which means, some software that has DTrace probes is now not
compiling on Lion.
> 
> Another difference is the addition of:
> 
> #if !defined(DTRACE_PROBES_DISABLED) || !DTRACE_PROBES_DISABLED
> 
> but I don''t think that is breaking anything.
> 
> I''ve attached both header files from the OS versions.  Does Apple
know of this?
> 
> thanks,
> 
> Brendan
> 
> -- 
> Brendan Gregg, Joyent                      http://dtrace.org/blogs/brendan
>
<lion.h><snowleopard.h>_______________________________________________
> dtrace-discuss mailing list
> dtrace-discuss at opensolaris.org
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.opensolaris.org/pipermail/dtrace-discuss/attachments/20110727/cea7b934/attachment.html>

Oh, is this a gcc issue?  Does xcode handle it ok?

Brendan

On Thu, Jul 28, 2011 at 5:07 AM, James McIlree <jmcilree at apple.com>
wrote:
>
> I''m not able to reproduce, what compiler are you using?
>
> James M
>
> On Jul 27, 2011, at 4:00 PM, Brendan Gregg wrote:
>
> G''Day,
>
> The way isenabled probes are compiled on Mac OS X Lion has changed in a way
> that dosen''t compile (at least, with some compilers). 
Here''s the test to
> try:
>
> $ cat -n liontest.d
>      1 provider lion {
>      2 probe hello(const char *world);
>      3 };
> $ dtrace -h -s liontest.d -o liontest.h
>
> On Snow Leopard this output:
>
> $ cat -n liontest.h
> [...]
>     24 #define LION_HELLO_ENABLED() \
>     25 __dtrace_isenabled$lion$hello$v1()
>
> On Lion that is now:
>
> $ can -n liontest.h
> [...]
>     26 #define LION_HELLO_ENABLED() \
>     27        ({ int _r = __dtrace_isenabled$lion$hello$v1(); \
>     28                __asm__ volatile(""); \
>     29                _r; })
>
> Which doesn''t compile when used in if statements (the use case for
> isenabled).  Which means, some software that has DTrace probes is now not
> compiling on Lion.
>
> Another difference is the addition of:
>
> #if !defined(DTRACE_PROBES_DISABLED) || !DTRACE_PROBES_DISABLED
>
> but I don''t think that is breaking anything.
>
> I''ve attached both header files from the OS versions.  Does Apple
know of
> this?
>
> thanks,
>
> Brendan
>
> --
> Brendan Gregg, Joyent                      http://dtrace.org/blogs/brendan
> 
<lion.h><snowleopard.h>_______________________________________________
> dtrace-discuss mailing list
> dtrace-discuss at opensolaris.org
>
>
>

-- 
Brendan Gregg, Joyent                      http://dtrace.org/blogs/brendan
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.opensolaris.org/pipermail/dtrace-discuss/attachments/20110728/9e249cdf/attachment.html>

I tried llvm-gcc and clang, both worked for me. It looks like invoking gcc
"by name" gives you llvm-gcc on a default Xcode install.

	I think everything we ship should work, are you building your own gcc?

	James M


	
On Jul 27, 2011, at 11:44 PM, Brendan Gregg wrote:
> Oh, is this a gcc issue?  Does xcode handle it ok?
> 
> Brendan
> 
> On Thu, Jul 28, 2011 at 5:07 AM, James McIlree <jmcilree at
apple.com> wrote:
> 
> 	I''m not able to reproduce, what compiler are you using?
> 
> 	James M
> 
> On Jul 27, 2011, at 4:00 PM, Brendan Gregg wrote:
> 
>> G''Day,
>> 
>> The way isenabled probes are compiled on Mac OS X Lion has changed in a
way that dosen''t compile (at least, with some compilers). 
Here''s the test to try:
>> 
>> $ cat -n liontest.d 
>>      1	provider lion {
>>      2		probe hello(const char *world);
>>      3	};
>> $ dtrace -h -s liontest.d -o liontest.h
>> 
>> On Snow Leopard this output:
>> 
>> $ cat -n liontest.h
>> [...]
>>     24	#define	LION_HELLO_ENABLED() \
>>     25		__dtrace_isenabled$lion$hello$v1()
>> 
>> On Lion that is now:
>> 
>> $ can -n liontest.h
>> [...]
>>     26	#define LION_HELLO_ENABLED() \
>>     27	        ({ int _r = __dtrace_isenabled$lion$hello$v1(); \
>>     28	                __asm__ volatile(""); \
>>     29	                _r; })
>> 
>> Which doesn''t compile when used in if statements (the use case
for isenabled).  Which means, some software that has DTrace probes is now not
compiling on Lion.
>> 
>> Another difference is the addition of:
>> 
>> #if !defined(DTRACE_PROBES_DISABLED) || !DTRACE_PROBES_DISABLED
>> 
>> but I don''t think that is breaking anything.
>> 
>> I''ve attached both header files from the OS versions.  Does
Apple know of this?
>> 
>> thanks,
>> 
>> Brendan
>> 
>> -- 
>> Brendan Gregg, Joyent                     
http://dtrace.org/blogs/brendan
>>
<lion.h><snowleopard.h>_______________________________________________
>> dtrace-discuss mailing list
>> dtrace-discuss at opensolaris.org
> 
> 
> 
> 
> -- 
> Brendan Gregg, Joyent                      http://dtrace.org/blogs/brendan
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.opensolaris.org/pipermail/dtrace-discuss/attachments/20110728/08b87d45/attachment.html>