RE: [power-gnu-discuss] gcc option for e500v2

["Liu Yu-B13201" <B13201@xxxxxxxxxxxxx>]

> -----Original Message-----
> From: Nathan Sidwell [mailto:nathan@xxxxxxxxxxxxxxxx] 
> Sent: Friday, October 12, 2007 2:38 PM
> To: Liu Yu-B13201
> Cc: power-gnu-discuss@xxxxxxxxxxxxxxxx
> Subject: Re: [power-gnu-discuss] gcc option for e500v2
> 
> Liu Yu-B13201 wrote:
> > Hi all,
> > 
> > I am using the cs-released gcc-4.1.78 which configured for 
> e500v2. It 
> > has some special optimization options such as "-mcpu=8548 -mspe=yes 
> > -mabi=spe -mfloat-gprs=double". Can anyone tell me what effect they 
> > can make?
> 
> What is unclear in the manual about these options?  If you 
> have specific questions that are not answered there, we can 
> improve the documentation.
> 
> > In fact, if these optimization options are necessary while building 
> > package, I need to configure them set default.
> 
> The '-te500v2' option is sufficent to target the e500v2 core.
> 
> > BTW, option "-m64" is not support on my target. When use 
> this option, 
> > the error info is like below:
> > 
> >> powerpc-linux-gnuspe-gcc -m64  test.c
> > test.c:1: error: -m64 not supported in this configuration
> > 
> > Does anyone know how to enable it?
> 
> The toolchain does not target 64 bit systems.

Although the toolchain does not target 64 bit systems, I need to
configure it compatible with some classic powerpc script which contains
m64 option. So I don't care about what kind of code it will generate
when option m64 is set,  just passing the build will be OK.

I have tried to add rule %{m64: -m32} into gcc spec file, but gcc still
provided me this error info. Do I have to change the source code to meet
my demand?


> 
> nathan
> -- 
> Nathan Sidwell    ::   http://www.codesourcery.com   ::       
>   CodeSourcery
> 
>