[Asterisk-code-review] Test cel: Fails when DONT OPTIMIZE is off (asterisk[13])

[Chris Savinovich]

Chris Savinovich has posted comments on this change. ( https://gerrit.asterisk.org/10891 )

Change subject: Test_cel: Fails when DONT_OPTIMIZE is off
......................................................................


Patch Set 2:

> It seems wrong to do this.  I think we force inlining more than we
 > really should.  Inlined functions should be trivial since the point
 > of inlining them is to eliminate the function call overhead that
 > suppossedly dwarfs the execution time of the function itself.  We
 > should be letting the compiler decide if the function should be
 > inlined or not as it is in a position to really know if inlining is
 > justified.
 > 
 > That said, if there really is a bug in the compiler then what else
 > does it affect?  Why won't it bite outside the testsuite under
 > normal builds?

You know Richard, you gave me an idea: all other usages of functions using force_inline should also be checked to confirm if affected by the same bug.  Some comments in Google don't see this as a bug, they attribute it to wrong application of code, as you also suggest. Having said that, this specific fix is actually skipping the force_inline call.
I also added condition 6 to the list of parameters that cause this error ("6. There must exist a certain combination of threading") to support the idea that this "bug" may only occur only when all 6 conditions are present.


-- 
To view, visit https://gerrit.asterisk.org/10891
To unsubscribe, or for help writing mail filters, visit https://gerrit.asterisk.org/settings

Gerrit-Project: asterisk
Gerrit-Branch: 13
Gerrit-MessageType: comment
Gerrit-Change-Id: Ied32e0071f12ed9d5f3b4cdd878b2532a1c769d7
Gerrit-Change-Number: 10891
Gerrit-PatchSet: 2
Gerrit-Owner: Chris Savinovich <csavinovich at digium.com>
Gerrit-Reviewer: Chris Savinovich <csavinovich at digium.com>
Gerrit-Reviewer: Friendly Automation (1000185)
Gerrit-Reviewer: Richard Mudgett <rmudgett at digium.com>
Gerrit-Comment-Date: Sun, 20 Jan 2019 18:23:00 +0000
Gerrit-HasComments: No
Gerrit-HasLabels: No
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.digium.com/pipermail/asterisk-code-review/attachments/20190120/ac9ecea3/attachment.html>