RE: strange timing behaviour

["Amir Tsvitov" <amir dot tsvitov at percello dot com>]

Hello Yann,

Thanks for suggestions, I'm in the process of trying them.
BTW, I built a 4.2.1 gcc using the latest cross-ng tool and had some problems with the custom kernel headers we're using.
Our kernel is 2.6.28.4 and the asm include directory has changed in that kernel version. I noticed that when working with the standard headers the tool requires that you specify the kernel version. Perhaps even with custom headers this is still required so that the tool knows where to look for stuff.

Thanks again for the great tool,
Amir

> -----Original Message-----
> From: Yann E. MORIN [mailto:yann.morin.1998@anciens.enib.fr]
> Sent: Wednesday, July 22, 2009 8:55 PM
> To: crossgcc@sourceware.org
> Cc: Amir Tsvitov
> Subject: Re: strange timing behaviour
> 
> Hell Amir!
> All,
> 
> On Wednesday 22 July 2009 09:56:03 Amir Tsvitov wrote:
> > The toolchain worked great until just recently when we discovered
> that
> > when we move a certain function around to different files in the
> source
> > tree we get large delays (~few milli seconds) between the call to the
> > function and its execution.? When the function is placed in some
> files
> > there is no problem while in other files there is the large delay.
> 
> And what's your program doing in the meantime: is it running, or is
> it sleeping?
> 
> Then try to run it through gdb and stop just prior to calling the
> function.
> Then, look where it's jumping to: is it your function's body, or some
> other
> code?
> 
> If it's jumping elsewhere, then try to find where that code comes from.
> Maybe it's some C++ handling for whatever runtime.
> 
> If it's directly to your function's body, what's going on next? Is it
> directly executing your code, or calling soem preamble (ie another
> call)?
> 
> Also, how is your program linked? Are all your object files linked
> together, or are you building shared libraries with some of them? If
> you're building shared libs, then is your function going into one of
> them instead of the main executable?
> 
> > This only happen when we compile with g++, with gcc there is no
> problem
> > (our code is pure C but for various reasons we are required to
> compile
> > it with g++)
> > Any ideas as to what could be causing this and how to further
> investigate
> > it?
> 
> For one, I'd try to relax the requirement for g++ if you're doing pure
> C
> and gcc does work. It smells like C++ runtime overhead, to me...
> 
> Regards,
> Yann E. MORIN.
> 
> --
> .-----------------.--------------------.------------------.--------------------.
> |  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics'
> conspiracy: |
> | +0/33 662376056 | Software  Designer | \ / CAMPAIGN     |  ___
>         |
> | --==< ^_^ >==-- `------------.-------:  X  AGAINST      |  \e/  There
> is no  |
> | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v
> conspiracy.  |
> `------------------------------^-------^------------------^--------------------'
> 


--
For unsubscribe information see http://sourceware.org/lists.html#faq