On Thu, 28 Jan 1999, Evgeny Stambulchik wrote:
> > (just pedantic, but the right fix is to make it (char *) not (unsigned
> > char *), because this is how it was defined and used previously.
>
> Hmm, I also used signed at the beginning, but then looked into 2.2.0 sources and
> there it's unsi
MOLNAR Ingo <[EMAIL PROTECTED]> wrote:
> > -#define PARAM empty_zero_page
> > +#define PARAM ((unsigned char *)empty_zero_page)
>
> (just pedantic, but the right fix is to make it (char *) not (unsigned
> char *), because this is how it was defined and used previously.
Hmm, I also used s
On Thu, 28 Jan 1999, Evgeny Stambulchik wrote:
> Usually, asm/pgtable.h is not sourced in setup.c, but with the latest raid
> patch, AND if compiled for SMP, it is. Here is a fix (non-Intel archs may need
> something similar, too):
this is already fixed in the 19980128 RAID patch. (this should
Hello,
linux/arch/i386/kernel/setup.c defines
extern char empty_zero_page[PAGE_SIZE];
whereas in include/asm-i386/pgtable.h we see:
extern unsigned long empty_zero_page[1024];
Usually, asm/pgtable.h is not sourced in setup.c, but with the latest raid
patch, AND if compiled for SMP, it is. He