qemu fixes

Hi,

This is a series of four patches which improve support for qemu on NetBSD.

sys-queue.h defines _SYS_QUEUE_H_ which is also defined by the <sys/queue.h> system header. <sys/disk.h> uses SLIST_ENTRY on NetBSD, which doesn't exist in sys-queue.h. Therefore, include <sys/queue.h> before including sys-queue.h.

AMD Saxony, Dresden, Germany
Operating System Research Center

Legal Information:
AMD Saxony Limited Liability Company & Co. KG
Sitz (Geschäftsanschrift):
   Wilschdorfer Landstr. 101, 01109 Dresden, Deutschland
Registergericht Dresden: HRA 4896
vertretungsberechtigter Komplementär:
   AMD Saxony LLC (Sitz Wilmington, Delaware, USA)
Geschäftsführer der AMD Saxony LLC:
   Dr. Hans-R. Deppe, Thomas McCoy

Index: block.c
===================================================================
@@ -21,6 +21,11 @@
  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
  * THE SOFTWARE.
  */
+#include "config-host.h"
+#ifdef _BSD	/* include native header before sys-queue.h */
+#include <sys/queue.h>
+#endif
+
 #include "qemu-common.h"
 #include "console.h"
 #include "block_int.h"

@@ -29,7 +34,6 @@
 #include <sys/types.h>
 #include <sys/stat.h>
 #include <sys/ioctl.h>
-#include <sys/queue.h>
 #include <sys/disk.h>
 #endif
 

Does this patch (which my mail client unfortunately refuses to cite...) take into account that we now have TAILQ_FOREACH_SAFE which NetBSD may lack?

Jan

Siemens AG, Corporate Technology, CT SE 2 ES-OS
Corporate Competence Center Embedded Linux

Yes, NetBSD has it:

#define TAILQ_FOREACH_SAFE(var, head, field, next)                      \
        for ((var) = ((head)->tqh_first);                               \
                (var) != NULL && ((next) = TAILQ_NEXT(var, field), 1);  \
                (var) = (next))

Another option to this patch would be:

Renaming _SYS_QUEUE_H_ to QEMU_SYS_QUEUE_H_
at least allows to include both sys-queue.h and <sys/queue.h>.

If the macros are not compatible, gcc prints a warning about redefining them. Using the -Wredundant-decls warning flag, gcc always prints a warning when you try override or duplicate macro defines and function declarations.

Christoph

AMD Saxony, Dresden, Germany
Operating System Research Center

Legal Information:
AMD Saxony Limited Liability Company & Co. KG
Sitz (Geschäftsanschrift):
   Wilschdorfer Landstr. 101, 01109 Dresden, Deutschland
Registergericht Dresden: HRA 4896
vertretungsberechtigter Komplementär:
   AMD Saxony LLC (Sitz Wilmington, Delaware, USA)
Geschäftsführer der AMD Saxony LLC:
   Dr. Hans-R. Deppe, Thomas McCoy

I don't think allowing inclusion of both headers in their current form is a good idea. But we may try harder to make the best out of the host's sys/queue.h and provide missing support in form of

#ifndef MACRO_XYZ
#define MACRO_XYZ ...
#endif

Jan

Siemens AG, Corporate Technology, CT SE 2 ES-OS
Corporate Competence Center Embedded Linux

In message: <[email protected]>
Christoph Egger <[email protected]> writes:
: On Friday 28 November 2008 12:34:54 Jan Kiszka wrote:
: > Christoph Egger wrote:
: > > Hi,
: > >
: > > This is a series of four patches which improve support
: > > for qemu on NetBSD.
: > >
: > > sys-queue.h defines _SYS_QUEUE_H_ which is also defined by : > > the <sys/queue.h> system header. <sys/disk.h> uses SLIST_ENTRY : > > on NetBSD, which doesn't exist in sys-queue.h. Therefore, : > > include <sys/queue.h> before including sys-queue.h.
: >
: > Does this patch (which my mail client unfortunately refuses to cite...) : > take into account that we now have TAILQ_FOREACH_SAFE which NetBSD may : > lack?
:
: Yes, NetBSD has it:
:

: #define TAILQ_FOREACH_SAFE(var, head, field, next)                      \
:         for ((var) = ((head)->tqh_first);                               \
:                 (var) != NULL && ((next) = TAILQ_NEXT(var, field), 1);  \

: (var) = (next))
:
: Another option to this patch would be:
:
: Renaming _SYS_QUEUE_H_ to QEMU_SYS_QUEUE_H_
: at least allows to include both sys-queue.h and <sys/queue.h>.

That would be the better option, because _XXX is reserved for the compiler/library/toolchain, not for the application...

Warner

Thanks, applied.