[DTrace-devel] [PATCH v3] dtrace provider: add a predicate against the current tgid

[Nick Alcock]

On 29 Feb 2024, Eugene Loh spake thusly:

> On 2/29/24 14:47, Nick Alcock wrote:
>
>> On 29 Feb 2024, Kris Van Hees verbalised:
>>
>>> On Thu, Feb 29, 2024 at 05:42:30PM +0000, Nick Alcock via DTrace-devel wrote:
>>>> Signed-off-by: Nick Alcock <nick.alcock at oracle.com>
>>> Reviewed-by: Kris Van Hees <kris.van.hees at oracle.com>
>> This commit induces a failure of
>> test/unittest/dtrace-util/tst.DisOption.sh, but this test is testing an
>> undocumented option, is carrying out transformations with awk etc on it,
>> that are uncommented and entirely opaque to me (it seems to be splitting
>> things up then diffing the original against the split-up pieces and
>> expecting it to work?), and honestly I think Eugene can probably fix it
>> in seconds. I don't have any idea what it's trying to verify. A comment
>> or two might be a good idea.
>
> I'll add some comments.
>
> Anyhow, it's trying to confirm that the option values dump the appropriate phases.  So it runs a few times for different phases and
> checks for consistency.
>
> One question:  will the disasm output be the same from run to run? BOOTTM will change, so the test filters that out.
>
> With this patch, the tgid value in the predicate also changes.  So that should also be filtered.

Ahhhhh, I see! Of course it does, I completely failed to notice that.
Dead obvious now you mention it.

(The real problem was that the tgid was so high I didn't recognize it as
anything like a PID. echo $$ would soon have taught me otherwise.)

I'll give it a try tomorrow morning.

-- 
NULL && (void)