Keith
Thanks for your detailed response. Your post resulted in me looking at
kernel.org kernels. I find kernel.org kernels do not need this change.
This change is required for the "kexec" patches (which was merged ahead
of our kdb patches in our tree). Anyone doing a merge with kexec will
want this patch.
Sorry for the confusion.
regards
-steve
On Mon, 2005-06-06 at 19:27, Keith Owens wrote:
> On Mon, 06 Jun 2005 14:38:11 -0700,
> Steven Dake <[EMAIL PROTECTED]> wrote:
> >I couldn't get kdb to compile with KDB modules set to * (but m works).
>
> I build kdb for i386 with gcc 3.4.3 all the time (FC3). It has no
> problems with any value of KDB_MODULES, I just tried y, m and n, all
> built clean. '*' is not a valid setting for a config tristate
> variable.
>
> What value are you using for KDB_MODULES? What tool chain errors are
> you seeing?
>
> > __con_initcall_end = .;
> > __kdb_initcall_start = .;
> >- .kdb_initcall.init : { *(.kdb_initcall.init) }
> >+ .kdb_initcall.init : AT(ADDR(.kdb_initcall.init) - LOAD_OFFSET) {
> >+ *(.kdb_initcall.init)
> >+ }
>
> .kdb_initcall.init is functionally identical to the immediately
> preceding .con_initcall.init, and to all the other init sections in
> arch/i386/vmlinux.lds.S. If kdb needs changing then they all do.
>
---------------------------
Use http://oss.sgi.com/ecartis to modify your settings or to unsubscribe.
