On Wed, Mar 16, 2011 at 05:12:17PM +0100, Alexander Graf wrote:
> On 03/16/2011 05:56 AM, David Gibson wrote:
[snip]
> >+#include "hw.h"
> >+#include "net.h"
> >+#include "hw/qdev.h"
> >+#include "hw/spapr.h"
> >+#include "hw/spapr_vio.h"
> >+
> >+#include<libfdt.h>
>
> Hrm - might be good to protect compilation against existence of fdt
> then?
Ah, yeah, you've reminded me. So this appears in a number of places.
We basically have to have fdt support for the pseries platform.
What's the right way to make the whole machine definition conditional
upon libfdt?
Incidentally, can I suggest it might be a good idea to include libfdt
in the qemu tree and make it always-on instead of configurable.
[snip]
> >+ lbuf = qemu_mallocz(total_len);
>
> Do you really need the zeroing here? In fact, this looks like a good
> candidate for alloca :).
Ah, good idea.
[snip]
> >-#define SPAPR_VIO_TCE_PAGE_SHIFT 12
> >-#define SPAPR_VIO_TCE_PAGE_SIZE (1ULL<<
> >SPAPR_VIO_TCE_PAGE_SHIFT)
> >-#define SPAPR_VIO_TCE_PAGE_MASK (SPAPR_VIO_TCE_PAGE_SIZE - 1)
> >+#define SPAPR_VIO_TCE_PAGE_SHIFT 12
> >+#define SPAPR_VIO_TCE_PAGE_SIZE (1ULL<< SPAPR_VIO_TCE_PAGE_SHIFT)
> >+#define SPAPR_VIO_TCE_PAGE_MASK (SPAPR_VIO_TCE_PAGE_SIZE - 1)
>
> Those shouldn't have been tabs in the first place :)
Fixed in the original patch now.
--
David Gibson | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_
| _way_ _around_!
http://www.ozlabs.org/~dgibson
