Re: [Xen-devel] [PATCH v3 4/5] x86: Port the basic alternative mechanism from Linux to Xen

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 03.06.14 at 13:42, <feng.wu@xxxxxxxxx> wrote:
> I am a little curious, from xen.lds.S, we can see both *(.init.data) and 
> *(.init.rodata) go into section .init.data.
> So even we use __initconst and "const" modifier, the data is still not 
> constant, right?

The use of the "const" modifier makes it so the compiler guarantees
const-ness. But yes, this isn't being enforced at the processor level.

The same compiler issue as observed here prevents the use of
__initdata for "const" objects: Some older gcc versions produce
"section type conflict" warnings/errors if you try to put both const
and non-const data into the same section.

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel