[Qemu-devel] [PATCH v2] configure: remove duplicate test

[Qemu-devel] [PATCH v2] configure: remove duplicate test

[Paolo Bonzini]

We already had a test to add -march=i486 when needed.  Make the
existing test independent of vhost-net, so that it is also used
under Win32.  Also move the test earlier to avoid mucking with
QEMU_CFLAGS too late in the configure process.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
	v1->v2: move test early

 configure | 56 ++++++++++++++++++++++++++------------------------------
 1 file changed, 26 insertions(+), 30 deletions(-)

diff --git a/configure b/configure
index b12d9ac..b2e0804 100755
--- a/configure
+++ b/configure
@@ -571,11 +571,6 @@ if test "$mingw32" = "yes" ; then
   QEMU_CFLAGS="-DWIN32_LEAN_AND_MEAN -DWINVER=0x501 $QEMU_CFLAGS"
   # enable C99/POSIX format strings (needs mingw32-runtime 3.15 or later)
   QEMU_CFLAGS="-D__USE_MINGW_ANSI_STDIO=1 $QEMU_CFLAGS"
-  if test "$cpu" = "i386"; then
-    # We need something better than i386 for __sync_val_compare_and_swap
-    # and can expect that QEMU will only run on i686 or later.
-    QEMU_CFLAGS="-march=i686 $QEMU_CFLAGS"
-  fi
   LIBS="-lwinmm -lws2_32 -liphlpapi $LIBS"
 cat > $TMPC << EOF
 int main(void) { return 0; }
@@ -1308,9 +1303,33 @@ EOF
   fi
 fi
 
-#
+##########################################
+# __sync_fetch_and_and requires at least -march=i486. Many toolchains
+# use i686 as default anyway, but for those that don't, an explicit
+# specification is necessary
+
+if test "$cpu" = "i386"; then
+  cat > $TMPC << EOF
+static int sfaa(int *ptr)
+{
+  return __sync_fetch_and_and(ptr, 0);
+}
+
+int main(void)
+{
+  int val = 42;
+  sfaa(&val);
+  return val;
+}
+EOF
+  if ! compile_prog "" "" ; then
+    QEMU_CFLAGS="-march=i486 $QEMU_CFLAGS"
+  fi
+fi
+
+#########################################
 # Solaris specific configure tool chain decisions
-#
+
 if test "$solaris" = "yes" ; then
   if has $install; then
     :
@@ -3047,29 +3066,6 @@ if test "$trace_backend" = "dtrace"; then
 fi
 
 ##########################################
-# __sync_fetch_and_and requires at least -march=i486. Many toolchains
-# use i686 as default anyway, but for those that don't, an explicit
-# specification is necessary
-if test "$vhost_net" = "yes" && test "$cpu" = "i386"; then
-  cat > $TMPC << EOF
-static int sfaa(int *ptr)
-{
-  return __sync_fetch_and_and(ptr, 0);
-}
-
-int main(void)
-{
-  int val = 42;
-  sfaa(&val);
-  return val;
-}
-EOF
-  if ! compile_prog "" "" ; then
-    QEMU_CFLAGS="-march=i486 $QEMU_CFLAGS"
-  fi
-fi
-
-##########################################
 # check and set a backend for coroutine
 
 # We prefer ucontext, but it's not always possible. The fallback
-- 
1.8.1.4

Re: [Qemu-devel] [PATCH v2] configure: remove duplicate test

[Peter Maydell]

On 9 April 2013 14:17, Paolo Bonzini <pbonzini@redhat.com> wrote:
> We already had a test to add -march=i486 when needed.  Make the
> existing test independent of vhost-net, so that it is also used
> under Win32.  Also move the test earlier to avoid mucking with
> QEMU_CFLAGS too late in the configure process.
>
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Do we care about passing this flag through to submodules?
If so we probably need to do something with EXTRA_CFLAGS
(cf Peter Crosthwaite's DTC submodule patches).

If not, this looks OK.

thanks
-- PMM

Re: [Qemu-devel] [PATCH v2] configure: remove duplicate test

[Paolo Bonzini]

Il 09/04/2013 19:22, Peter Maydell ha scritto:
> On 9 April 2013 14:17, Paolo Bonzini <pbonzini@redhat.com> wrote:
>> We already had a test to add -march=i486 when needed.  Make the
>> existing test independent of vhost-net, so that it is also used
>> under Win32.  Also move the test earlier to avoid mucking with
>> QEMU_CFLAGS too late in the configure process.
>>
>> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> 
> Do we care about passing this flag through to submodules?

I don't think so, submodules would have their own configure tests.

Paolo

> If so we probably need to do something with EXTRA_CFLAGS
> (cf Peter Crosthwaite's DTC submodule patches).
> 
> If not, this looks OK.
> 
> thanks
> -- PMM
> 
>