This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Default values of "struct item *items[MAX_ITEMS]" array?
- From: Daniel Jacobowitz <drow at false dot org>
- To: Joe Buck <Joe dot Buck at synopsys dot COM>
- Cc: Andriy Korud <a dot korud at vector dot com dot pl>,Richard Guenther <rguenth at tat dot physik dot uni-tuebingen dot de>,gcc at gcc dot gnu dot org
- Date: Sun, 16 May 2004 23:49:10 -0400
- Subject: Re: Default values of "struct item *items[MAX_ITEMS]" array?
- References: <80908CC5B2C9DB47AAF8C77892FCB44315F771@lion.vector.com.pl> <20040514085527.A28677@synopsys.com>
On Fri, May 14, 2004 at 08:55:27AM -0700, Joe Buck wrote:
> On Fri, May 14, 2004 at 10:54:38AM +0200, Andriy Korud wrote:
> > It seems to be compiler/optimization problem - the same kernel compiled with 2.95.4 works fine, however code compiles with 3.x fails.
> > Maybe compiler optimization just throws away kernel bss zeroing code?
>
> There is no "bss zeroing code" generated by either compiler.
> The problem must lie elsewhere.
FYI, the most frequent cause of "my .bss didn't get zeroed" reports is
when a newer compiler generates .sbss instead, confusing the boot
loader / module loader / dynamic linker / whatever for the target
platform. That's what this sounds like.
--
Daniel Jacobowitz