Re: [Open64-devel] disabling the compiler generated prologue/epilogue

[jiaqi z. <zat...@gm...>]

It looks as if it has something to do with the .debug_frame section. I
removed the -g switch and it works now. However, the debug info is
very important that I do not want to sacrifice it. As I am not
familiar with the cg part, is there any suggestion that I may revert
this info? Or should I use other mechanisms?

On Fri, Apr 17, 2009 at 9:28 AM, jiaqi zhang <zat...@gm...> wrote:
> Thanks, Robert. I tried the switch, but it doesn't seem to work. And
> also, I just need to omit the frame operations for certain functions,
> e.g. marked by specific directives.
>
>
> On Fri, Apr 17, 2009 at 2:00 AM, Robert Hundt <rob...@gm...> wrote:
>> Does -fomit-frame-pointer work?
>>   R
>>
>> On Thu, Apr 16, 2009 at 12:33 AM, jiaqi zhang <zat...@gm...> wrote:
>>>
>>> To be more specific, I want convert the functions from
>>>
>>> func1:
>>>       push  %ebp
>>>       mov   %esp, %ebp
>>>       ......
>>>       leave
>>>       ret
>>>
>>> to
>>>
>>> func1:
>>>       ......
>>>
>>>
>>> Thank you very much.
>>>
>>>
>>> On Thu, Apr 16, 2009 at 2:54 PM, jiaqi zhang <zat...@gm...> wrote:
>>> > Hi all,
>>> >
>>> > For certain reasons, I want to remove the compiler generated stack
>>> > operations (such as push %ebp, ret....) for specific functions. It
>>> > seems to be similar with the "naked" flag in gcc. I tried to make
>>> > these functions SMODEL_SMALL when choosing stack models. However, the
>>> > assembler complains unresolved expressions in the .LCIE section. What
>>> > should I do? Or is there any thing in Open64 I can refer to?
>>> >
>>> > Any suggestion is appreciated.
>>> >
>>> >
>>> >
>>> > --
>>> > Best regards
>>> > JiaQi Zhang
>>> >
>>>
>>>
>>>
>>> --
>>> Best regards
>>> JiaQi Zhang
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Stay on top of everything new and different, both inside and
>>> around Java (TM) technology - register by April 22, and save
>>> $200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco.
>>> 300 plus technical and hands-on sessions. Register today.
>>> Use priority code J9JMT32. http://p.sf.net/sfu/p
>>> _______________________________________________
>>> Open64-devel mailing list
>>> Ope...@li...
>>> https://lists.sourceforge.net/lists/listinfo/open64-devel
>>
>>
>>
>> --
>> Robert Hundt
>>
>
>
>
> --
> Best regards
> JiaQi Zhang
>



-- 
Best regards
JiaQi Zhang
High Performance Computing Institute,
Tsinghua University