[DTrace-devel] [PATCH 13/38] Hide dtrace_actdesc_t until it is needed

[Eugene Loh]

On 7/18/24 16:02, Kris Van Hees wrote:

> On Thu, Jun 27, 2024 at 01:34:30AM -0400, eugene.loh at oracle.com wrote:
>> From: Eugene Loh <eugene.loh at oracle.com>
>>
>> Signed-off-by: Eugene Loh <eugene.loh at oracle.com>
> I think you should just get rid of it.  Worst case, we can always revert the
> patch.  But this type is only used in code that we do not plan to revive since
> we no longer do DOF the same way as the older version.  Even if we were to
> re-implement the functionality to save compiled probe programs to file, it
> would look very different and certainly not be action-based but rather BPF
> code based.

It would seem to me that the declaration (this patch) and the code that 
uses it should go together.  So, are you saying the inoperative code 
that references this declaration should also be removed?  It would seem 
funny to me to remove the declaration (this patch) while (admittedly 
inoperative) code that references it is left alone.

>
>> ---
>>   include/dtrace/enabling.h | 3 +++
>>   1 file changed, 3 insertions(+)
>>
>> diff --git a/include/dtrace/enabling.h b/include/dtrace/enabling.h
>> index f1ec444c..2562c87b 100644
>> --- a/include/dtrace/enabling.h
>> +++ b/include/dtrace/enabling.h
>> @@ -43,6 +43,8 @@ typedef struct dtrace_probedesc {
>>   	const char	*prb;			/* probe name */
>>   } dtrace_probedesc_t;
>>   
>> +#ifdef FIXME
>> +This type is used only in #ifdef FIXME code.
>>   typedef struct dtrace_actdesc {
>>   	struct dtrace_difo *dtad_difo;		/* pointer to DIF object */
>>   	dtrace_actkind_t dtad_kind;		/* kind of action */
>> @@ -50,6 +52,7 @@ typedef struct dtrace_actdesc {
>>   	uint64_t dtad_arg;			/* action argument */
>>   	uint64_t dtad_uarg;			/* user argument */
>>   } dtrace_actdesc_t;
>> +#endif
>>   
>>   typedef struct dtrace_ecbdesc {
>>   	dtrace_probedesc_t dted_probe;		/* probe description */
>> -- 
>> 2.18.4
>>