Messages in this thread | | | From | Keith Owens <> | Subject | Re: PCI Express support for 2.4 kernel | Date | Tue, 16 Dec 2003 05:16:37 +1100 |
| |
On Mon, 15 Dec 2003 18:38:16 +0100, Tomas Szepe <szepe@pinerecords.com> wrote: >And w/ gcc-3.3.2, you get what you'd actually expect :) > >gcc version 3.3.2 >$ cat x.c >static int a1 = 0; >static int a2 = 1; >static int a3; >$ gcc -S x.c >$ cat x.s > .file "x.c" > .local a1 > .comm a1,4,4 > .data > .align 4 > .type a2, @object > .size a2, 4 >a2: > .long 1 > .local a3 > .comm a3,4,4 > .ident "GCC: (GNU) 3.3.2"
Does gcc 3.3.2 handle sections correctly when it optimizes zero assignments to use bss? With just this line in x.c
static int __attribute__ ((__section__ (".init.data"))) a1 = 0;
what does objdump -h report? If bss is 4 bytes then gcc 3.3.2 is breaking the kernel's use of init data. If bss is 0 and .init.data is 4 then we are OK.
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |