Re: [patch] Prevent dead code/data removal with gcc 3.4

[Ralf Baechle <ralf@linux-mips.org>]

On Sat, Feb 14, 2004 at 02:28:01AM +0100, Thiemo Seufer wrote:

> > It's the gcc generated function epilogue which is the problem.  There's
> > no reliable way to work around that ...
> 
> ITYM prologue. It has to follow the ABI specification, so $fp is the only
> possibly problematic one, and that's excluded by -fomit-frame-pointers.

Daniel Jacobowitz reported the problem which indeed was about $30.  Since
the kernel uses -fomit-frame-pointer by default (and -O1 enables it
by default on MIPS anyway) I would assume his kernel was built using that
option.  Maybe he can elaborate ...

Anyway, gcc could load next weeks lucky lottery numbers into the
s-registers after saving them.  That'd break save_static but not the
ABI which only promises to restore the old values in s-registers on
return.

  Ralf