[Buildroot] build of gcc 4.3.5 broken with BR2_GCC_USE_SJLJ_EXCEPTIONS

[Buildroot] build of gcc 4.3.5 broken with BR2_GCC_USE_SJLJ_EXCEPTIONS

[Ossy]

Hi mailinglist,

I don't need this particular option (BR2_GCC_USE_SJLJ_EXCEPTIONS) but 
anyway my buildroot version (current git tree) wasn't able to build 
4.3.5-final with BR2_GCC_USE_SJLJ_EXCEPTIONS=y. After deactivating it 
everything was fine again.

Maybe someone is able to reproduce this behaviour and file a bug.

Regards,
Marcus

[Buildroot] build of gcc 4.3.5 broken with BR2_GCC_USE_SJLJ_EXCEPTIONS

[Yann E. MORIN]

Marcus, All,

On Thursday 22 July 2010 23:59:25 Ossy wrote:
> I don't need this particular option (BR2_GCC_USE_SJLJ_EXCEPTIONS) but 
> anyway my buildroot version (current git tree) wasn't able to build 
> 4.3.5-final with BR2_GCC_USE_SJLJ_EXCEPTIONS=y. After deactivating it 
> everything was fine again.

Yes, it happens too for me with armeb. Other archs are fine for me with
sjlj enabled.

I did not dig up why, so I /simply/ turn off sjlj for ermeb.

Regards,
Yann E. MORIN.

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'

[Buildroot] build of gcc 4.3.5 broken with BR2_GCC_USE_SJLJ_EXCEPTIONS

[Thomas Petazzoni]

On Fri, 23 Jul 2010 00:05:22 +0200
"Yann E. MORIN" <yann.morin.1998@anciens.enib.fr> wrote:

> Marcus, All,
> 
> On Thursday 22 July 2010 23:59:25 Ossy wrote:
> > I don't need this particular option (BR2_GCC_USE_SJLJ_EXCEPTIONS)
> > but anyway my buildroot version (current git tree) wasn't able to
> > build 4.3.5-final with BR2_GCC_USE_SJLJ_EXCEPTIONS=y. After
> > deactivating it everything was fine again.
> 
> Yes, it happens too for me with armeb. Other archs are fine for me
> with sjlj enabled.
> 
> I did not dig up why, so I /simply/ turn off sjlj for ermeb.

Does having this BR2_GCC_USE_SJLJ_EXCEPTIONS option makes sense ? In
which case would this be useful ?

According to http://choices.cs.uiuc.edu/exceptions.pdf, the default
mechanism used by g++ for exceptions is faster and better than SJLJ.

If there are only corner use cases, I'm all for getting rid of this
option in BR. Advanced users can always use
BR2_EXTRA_GCC_CONFIG_OPTIONS to add --enable-sjlj-exceptions to the gcc
configuration if they want.

Thoughts ?

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

[Buildroot] build of gcc 4.3.5 broken with BR2_GCC_USE_SJLJ_EXCEPTIONS

[Khem Raj]

On (27/07/10 20:30), Thomas Petazzoni wrote:
> On Fri, 23 Jul 2010 00:05:22 +0200
> "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr> wrote:
> 
> > Marcus, All,
> > 
> > On Thursday 22 July 2010 23:59:25 Ossy wrote:
> > > I don't need this particular option (BR2_GCC_USE_SJLJ_EXCEPTIONS)
> > > but anyway my buildroot version (current git tree) wasn't able to
> > > build 4.3.5-final with BR2_GCC_USE_SJLJ_EXCEPTIONS=y. After
> > > deactivating it everything was fine again.
> > 
> > Yes, it happens too for me with armeb. Other archs are fine for me
> > with sjlj enabled.
> > 
> > I did not dig up why, so I /simply/ turn off sjlj for ermeb.
> 
> Does having this BR2_GCC_USE_SJLJ_EXCEPTIONS option makes sense ? In
> which case would this be useful ?

It makes sense if you are using older compilers and when DWARF unwinding
did not work completely. But otherwise its slow in performance and probably
can be replaced with dwarf2 exceptions I think hpux is one which does not
support dwarf2 exception per gcc sources.
sjlj handles exceptions faster but also adds burden of calling setjmp in 
non excepting case therefore slowing down the execution but code is
smaller.

I think you could leave this option in but dont make it default.

> 
> According to http://choices.cs.uiuc.edu/exceptions.pdf, the default
> mechanism used by g++ for exceptions is faster and better than SJLJ.
> 
> If there are only corner use cases, I'm all for getting rid of this
> option in BR. Advanced users can always use
> BR2_EXTRA_GCC_CONFIG_OPTIONS to add --enable-sjlj-exceptions to the gcc
> configuration if they want.
> 
> Thoughts ?
> 
> Thomas
> -- 
> Thomas Petazzoni, Free Electrons
> Kernel, drivers, real-time and embedded Linux
> development, consulting, training and support.
> http://free-electrons.com
> _______________________________________________
> buildroot mailing list
> buildroot at busybox.net
> http://lists.busybox.net/mailman/listinfo/buildroot