Skip to contentSkip to conversation

Win32 build

PATCH In review Filip Navara 3 participants 4 messages Jan 9, 2005 – Jan 11, 2005

lore
Share & tools
lore

Filip Navara

author
#0 lore raw

PATCH 1/1 Win32 build

Changes:
- Add x86_64-softmmu target to the build.
- Fix compiling with newer w32api headers.

Index: configure
===================================================================
RCS file: /cvsroot/qemu/qemu/configure,v
retrieving revision 1.50

configure +1 -1
diff -u -r1.50 configure
@@ -181,7 +181,7 @@
 strip="${cross_prefix}${strip}"
 
 if test "$mingw32" = "yes" ; then
-    target_list="i386-softmmu ppc-softmmu sparc-softmmu"
+    target_list="i386-softmmu ppc-softmmu sparc-softmmu x86_64-softmmu"
     EXESUF=".exe"
     gdbstub="no"
     oss="no"

Index: slirp/slirp.h
===================================================================
RCS file: /cvsroot/qemu/qemu/slirp/slirp.h,v
retrieving revision 1.9

slirp.h +1 -0
diff -u -r1.9 slirp.h
@@ -20,6 +20,7 @@
 typedef uint64_t u_int64_t;
 typedef char *caddr_t;
 
+# include <windows.h>
 # include <winsock2.h>
 # include <sys/timeb.h>
 # include <iphlpapi.h>

Index: target-i386/ops_template.h
===================================================================
RCS file: /cvsroot/qemu/qemu/target-i386/ops_template.h,v retrieving revision 1.8

ops_template.h +1 -1
diff -u -r1.8 ops_template.h
@@ -46,7 +46,7 @@
 #define SUFFIX q
 #define DATA_TYPE uint64_t
 #define DATA_STYPE int64_t
-#define DATA_MASK 0xffffffffffffffff
+#define DATA_MASK 0xffffffffffffffffULL
 #else
 #error unhandled operand size
 #endif
#1 lore raw

Le Sun, 09 Jan 2005 03:25:43 +0100, Filip Navara a écrit :

Quoted 11 lines

Changes:
- Add x86_64-softmmu target to the build. - Fix compiling with newer w32api headers.
Index: configure =================================================================== RCS
file: /cvsroot/qemu/qemu/configure,v retrieving revision 1.50 diff -u -r1.50 configure
--- configure 3 Jan 2005 23:38:40 -0000 1.50 +++ configure 9 Jan 2005
01:59:41 -0000 @@ -181,7 +181,7 @@
strip="${cross_prefix}${strip}"

Quoted 6 lines

if test "$mingw32" = "yes" ; then
- target_list="i386-softmmu ppc-softmmu sparc-softmmu" +
target_list="i386-softmmu ppc-softmmu sparc-softmmu x86_64-softmmu"

     EXESUF=".exe"
     gdbstub="no"
     oss="no"

Just a comment about qemu-system-x86_64, I have successfully loaded
http://bochs.sourceforge.net/guestos/linux.x86-64.bzImage on windows 98/750MHz duron. I went as far as the kernel mount the root ro and free unused memory (on linux too) because of tls.

Another thing which can be of interest: I have updated SDL to 1.2.8, wine (20041201) can now launch qemu.exe without setting bpp to a fixed value in sdl.c

qemu-doc.texi.new 2005-01-10 17:25:16.582909377 +0100 renamed from qemu-doc.texi 2004-12-20 01:18:55.000000000 +0100 +0 -3
@@ -1287,9 +1287,6 @@
 
 @end itemize
 
-Note: Currently, Wine does not seem able to launch
-QEMU for Win32.
-
 @section Mac OS X
 
 The Mac OS X patches are not fully merged in QEMU, so you should look
#2 lore raw
In reply to Filip Navara
Quoted 8 lines · Filip Navara

Index: target-i386/ops_template.h
===================================================================
RCS file: /cvsroot/qemu/qemu/target-i386/ops_template.h,v retrieving revision 1.8
diff -u -r1.8 ops_template.h

ops_template.h line 52

Quoted 8 lines
ops_template.h +1 -1
@@ -46,7 +46,7 @@
 #define SUFFIX q
 #define DATA_TYPE uint64_t
 #define DATA_STYPE int64_t
-#define DATA_MASK 0xffffffffffffffff
+#define DATA_MASK 0xffffffffffffffffULL
 #else
 #error unhandled operand size
 #endif

Is this really needed to compile ?

Fabrice.

author
#3 lore raw
Quoted 13 lines · Fabrice Bellard
Quoted 11 lines · Filip Navara

Index: target-i386/ops_template.h
===================================================================
RCS file: /cvsroot/qemu/qemu/target-i386/ops_template.h,v retrieving revision 1.8
diff -u -r1.8 ops_template.h

ops_template.h 9 Jan 2005 02:18:37 -0000 renamed from ops_template.h 3 Jan 2005 23:50:08 -0000 1.8 +0 -0
@@ -46,7 +46,7 @@
 #define SUFFIX q
 #define DATA_TYPE uint64_t
 #define DATA_STYPE int64_t
Quoted 8 lines

ops_template.h 9 Jan 2005 02:18:37 -0000 line 52

Quoted 5 lines
ops_template.h 9 Jan 2005 02:18:37 -0000 renamed from ops_template.h 3 Jan 2005 23:50:08 -0000 1.8 +1 -1
-#define DATA_MASK 0xffffffffffffffff
+#define DATA_MASK 0xffffffffffffffffULL
 #else
 #error unhandled operand size
 #endif

Is this really needed to compile ?

GCC reports lots of warnings like these otherwise:

qemu/target-i386/ops_template.h: In function `compute_all_incq':
qemu/target-i386/ops_template.h:182: warning: integer constant is too large for "long" type In file included from e:/DIFFs/qemu/target-i386/op.c:733:
qemu/target-i386/ops_template.h: In function `compute_all_decq':
qemu/target-i386/ops_template.h:204: warning: integer constant is too large for "long" type In file included from e:/DIFFs/qemu/target-i386/op.c:733:
qemu/target-i386/ops_template.h: In function `op_shrq_T0_T1':
qemu/target-i386/ops_template.h:436: warning: integer constant is too large for "long" type

- Filip