Re: subtle side effect of commit a1c48bb160f836

[Vineet Gupta]

Hi Michal,

On Thursday 18 June 2015 04:02 PM, Vineet Gupta wrote:
> On Thursday 18 June 2015 03:44 PM, Michal Marek wrote:
>> Dne 18.6.2015 v 10:45 Vineet Gupta napsal(a):
>>>> On Thursday 18 June 2015 01:43 PM, Michal Marek wrote:
>>>>>>>>>> Alternatively, as we already have CONFIG_CC_OPTIMIZE_FOR_SIZE,
>>>>>>>>>> a(nother) Kconfig option may make sense.
>>>>>> We can also introduce some ARCH_CFLAGS that is appended near the end of
>>>>>> the list, and have arc/Makefile add its -O3 there. But I'd like to why
>>>>>> the -O3 needs to be there in first place. 
>>>>
>>>> This is how historically ARC kernels have been built. We do track performance
>>>> results LMBench/hackbench... and going from -O3 to -O2 caused a sudden dip in some
>>>> of the numbers when we measured 3.18 (vs. 3.13)
>>>>
>>>>>> Obviously, the kernel works
>>>>>> with -O2, otherwise the regression would have been identified earlier.
>>>>
>>>> Its a performance thing - so yeah -O2 works, but -O3 works even better :-)
>>>>
>>>>>> So why can't users specify -O3 in KCFLAGS like on any other
>>>>>> architecture.
>>>>
>>>> Sweet, I didn't know about this. But I don't see any arch setting this - only tile
>>>> using it. So yeah - below does the trick !
>>>>
>>>> --------->
>>>> From 5e44cd2ed69b1d030b4cb87600d2767b69c35537 Mon Sep 17 00:00:00 2001
>>>> From: Vineet Gupta <vgupta@xxxxxxxxxxxx>
>>>> Date: Thu, 18 Jun 2015 13:54:01 +0530
>>>> Subject: [PATCH] ARC: Override toplevel default -O2 with -O3
>>>>
>>>> ARC kernels have historically been built with -O3, despite top level
>>>> Makefile defaulting to -O2. This was facilitated by implicitly ordering
>>>> of arch makefile include AFTER top level assigned -O2.
>>>>
>>>> An upstream fix to top level a1c48bb160f ("Makefile: Fix unrecognized
>>>> cross-compiler command line options") changed the ordering, making ARC
>>>> -O3 defunt.
>>>>
>>>> Fix that by NOT relying on any ordering whatsoever and use the right
>>>> mechanism to do the over-rides.
>>>>
>>>> Suggested-by: Michal Marek <mmarek@xxxxxxx>
>>>> Cc: Geert Uytterhoeven <geert@xxxxxxxxxxxxxx>
>>>> Cc: <stable@xxxxxxxxxxxxxxx>
>>>> Signed-off-by: Vineet Gupta <vgupta@xxxxxxxxxxxx>
>>>> ---
>>>>  arch/arc/Makefile | 1 +
>>>>  1 file changed, 1 insertion(+)
>>>>
>>>> diff --git a/arch/arc/Makefile b/arch/arc/Makefile
>>>> index 86c71b2089d2..c23f3f2b0ff8 100644
>>>> --- a/arch/arc/Makefile
>>>> +++ b/arch/arc/Makefile
>>>> @@ -44,6 +44,7 @@ endif
>>>>  ifndef CONFIG_CC_OPTIMIZE_FOR_SIZE
>>>>  # Generic build system uses -O2, we want -O3
>>>>  cflags-y  += -O3
>>>> +KCFLAGS += -O3
>>>>  endif
>> Uh, this is not what I meant. KCFLAGS is a *user* setting. It's meant to
>> be set in the environment or command line.
> 
> Well I don't want to rely on external settings whatsoever to enforce this. If this
> is not the right way, what do u suggest I do to help fix this.


Can I keep this seeming abuse of KCFLAGS or do u suggest alternate approach I can
code up !

Thx,
-Vineet
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/