Re: uname ?

[Erik Mouw]

On Wed, Apr 18, 2001 at 07:56:25PM +0200, Dragan Milenkovic wrote:
> <rpm-related>
> Few days ago, I found an incredibly usable option in latest gcc versions.
> gcc -march=k6 -mcpu=k6

That -mcpu flag is not necessary. From the gcc info files:

`-march=CPU TYPE'
     Generate instructions for the machine type CPU TYPE.  The choices
     for CPU TYPE are the same as for `-mcpu'.  Moreover, specifying
     `-march=CPU TYPE' implies `-mcpu=CPU TYPE'.

> I started rebuilding SRPMS with these as optflags, just to find out that
> some packages don't use optflags, but use -march=`uname -m`.

Which breaks very nice if you want to cross compile packages:

  erik@arthur:/tmp> arm-linux-gcc -march=`uname -m` -c foo.c
  cc1: bad value (i686) for -march= switch

> So ...
> </rpm-related>
> 
> What about uname -m and non-Intel processors? Is there going to be
> a change? I hope we all agree that K6 is not i586 or i686.
> I vote for breaking compatibility!

You want to fix the kernel when gcc is not even consistent across its
targets? Have a look at the gcc info files: the -march= flag for ARM
targets behaves completely different from its ix86 counterpart. And the
gcc Sparc target doesn't even recognise an -march= flag.

Erik

-- 
J.A.K. (Erik) Mouw, Information and Communication Theory Group, Department
of Electrical Engineering, Faculty of Information Technology and Systems,
Delft University of Technology, PO BOX 5031,  2600 GA Delft, The Netherlands
Phone: [phone]  Fax: [phone]  Email: [e-mail]
WWW: http://www-ict.its.tudelft.nl/~erik/