[Qemu-devel] [PATCH 0/3] Use glib 2.26 version macros

[Qemu-devel] [PATCH 0/3] Use glib 2.26 version macros

[marcandre.lureau]

From: Marc-André Lureau <marcandre.lureau@redhat.com>

Bump glib requirement to be able to use the GLIB_VERSION macros.

Marc-André Lureau (3):
  hw/acpi/aml-build: remove useless glib version check
  configure: require glib 2.26
  Remove glib < 2.26 compatibility code

 configure               | 5 +++--
 hw/acpi/aml-build.c     | 2 --
 include/glib-compat.h   | 4 ----
 tests/vhost-user-test.c | 4 ----
 4 files changed, 3 insertions(+), 12 deletions(-)

-- 
2.4.3

[Qemu-devel] [PATCH 1/3] hw/acpi/aml-build: remove useless glib version check

[marcandre.lureau]

From: Marc-André Lureau <marcandre.lureau@redhat.com>

2.22 is the minimum version required

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 hw/acpi/aml-build.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/hw/acpi/aml-build.c b/hw/acpi/aml-build.c
index 0d4b324..a00a0ab 100644
--- a/hw/acpi/aml-build.c
+++ b/hw/acpi/aml-build.c
@@ -1163,9 +1163,7 @@ void *acpi_data_push(GArray *table_data, unsigned size)
 
 unsigned acpi_data_len(GArray *table)
 {
-#if GLIB_CHECK_VERSION(2, 22, 0)
     assert(g_array_get_element_size(table) == 1);
-#endif
     return table->len;
 }
 
-- 
2.4.3

[Qemu-devel] [PATCH 2/3] configure: require glib 2.26

[marcandre.lureau]

From: Marc-André Lureau <marcandre.lureau@redhat.com>

This allows to use the GLIB_VERSION macros that will help to ensure the
newer symbols are not being used (or with compatibility code).

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 configure | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/configure b/configure
index f08327e..1bc832f 100755
--- a/configure
+++ b/configure
@@ -2843,7 +2843,8 @@ fi
 ##########################################
 # glib support probe
 
-glib_req_ver=2.22
+glib_req_ver=2.26
+glib_ver=GLIB_VERSION_2_26
 glib_modules=gthread-2.0
 if test "$modules" = yes; then
     glib_modules="$glib_modules gmodule-2.0"
@@ -2853,7 +2854,7 @@ for i in $glib_modules; do
     if $pkg_config --atleast-version=$glib_req_ver $i; then
         glib_cflags=`$pkg_config --cflags $i`
         glib_libs=`$pkg_config --libs $i`
-        CFLAGS="$glib_cflags $CFLAGS"
+        CFLAGS="$glib_cflags $CFLAGS -DGLIB_VERSION_MIN_REQUIRED=$glib_ver -DGLIB_VERSION_MAX_ALLOWED=$glib_ver"
         LIBS="$glib_libs $LIBS"
         libs_qga="$glib_libs $libs_qga"
     else
-- 
2.4.3

[Qemu-devel] [PATCH 3/3] Remove glib < 2.26 compatibility code

[marcandre.lureau]

From: Marc-André Lureau <marcandre.lureau@redhat.com>

Since 2.26 is now the minimum.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
---
 include/glib-compat.h   | 4 ----
 tests/vhost-user-test.c | 4 ----
 2 files changed, 8 deletions(-)

diff --git a/include/glib-compat.h b/include/glib-compat.h
index 318e000..46a2bc5 100644
--- a/include/glib-compat.h
+++ b/include/glib-compat.h
@@ -19,10 +19,6 @@
 #include <glib.h>
 
 /* GLIB version compatibility flags */
-#if !GLIB_CHECK_VERSION(2, 26, 0)
-#define G_TIME_SPAN_SECOND              (G_GINT64_CONSTANT(1000000))
-#endif
-
 #if !GLIB_CHECK_VERSION(2, 28, 0)
 static inline gint64 qemu_g_get_monotonic_time(void)
 {
diff --git a/tests/vhost-user-test.c b/tests/vhost-user-test.c
index 56df5cc..9f2254d 100644
--- a/tests/vhost-user-test.c
+++ b/tests/vhost-user-test.c
@@ -21,10 +21,6 @@
 #include <qemu/sockets.h>
 
 /* GLIB version compatibility flags */
-#if !GLIB_CHECK_VERSION(2, 26, 0)
-#define G_TIME_SPAN_SECOND              (G_GINT64_CONSTANT(1000000))
-#endif
-
 #if GLIB_CHECK_VERSION(2, 28, 0)
 #define HAVE_MONOTONIC_TIME
 #endif
-- 
2.4.3

Re: [Qemu-devel] [PATCH 0/3] Use glib 2.26 version macros

[Cornelia Huck]

On Mon, 12 Oct 2015 12:54:56 +0200
marcandre.lureau@redhat.com wrote:

> From: Marc-André Lureau <marcandre.lureau@redhat.com>
> 
> Bump glib requirement to be able to use the GLIB_VERSION macros.
> 
> Marc-André Lureau (3):
>   hw/acpi/aml-build: remove useless glib version check
>   configure: require glib 2.26
>   Remove glib < 2.26 compatibility code
> 
>  configure               | 5 +++--
>  hw/acpi/aml-build.c     | 2 --
>  include/glib-compat.h   | 4 ----
>  tests/vhost-user-test.c | 4 ----
>  4 files changed, 3 insertions(+), 12 deletions(-)
> 

Apologies if I missed some discussions, but isn't SLES 11 still at
2.22? (The SLES 11 system I used has been upgraded to SLES 12, so I
can't check myself.)

Do we still aim to be able to build on any still-supported enterprise
distro?

Re: [Qemu-devel] [PATCH 0/3] Use glib 2.26 version macros

[Daniel P. Berrange]

On Mon, Oct 12, 2015 at 12:54:56PM +0200, marcandre.lureau@redhat.com wrote:
> From: Marc-André Lureau <marcandre.lureau@redhat.com>
> 
> Bump glib requirement to be able to use the GLIB_VERSION macros.

We had a discussion about glib min version a few months back where it
was decided 2.22 was the maximum we could use at the current time
without us loosing support for certain platforms we care about.

Simply being able to use the GLIB_VERSION macro does not seem like a
compelling enough reason to re-vist this min version again right now,
as I doubt we've changed our mind on what platforms we wish to target
in the last couple of months.

Regards,
Daniel
-- 
|: http://berrange.com      -o-    http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org              -o-             http://virt-manager.org :|
|: http://autobuild.org       -o-         http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org       -o-       http://live.gnome.org/gtk-vnc :|

Re: [Qemu-devel] [PATCH 0/3] Use glib 2.26 version macros

[Marc-André Lureau]

Hi

On Mon, Oct 12, 2015 at 1:26 PM, Daniel P. Berrange <berrange@redhat.com> wrote:
> On Mon, Oct 12, 2015 at 12:54:56PM +0200, marcandre.lureau@redhat.com wrote:
>> From: Marc-André Lureau <marcandre.lureau@redhat.com>
>>
>> Bump glib requirement to be able to use the GLIB_VERSION macros.
>
> We had a discussion about glib min version a few months back where it
> was decided 2.22 was the maximum we could use at the current time
> without us loosing support for certain platforms we care about.
>
> Simply being able to use the GLIB_VERSION macro does not seem like a
> compelling enough reason to re-vist this min version again right now,
> as I doubt we've changed our mind on what platforms we wish to target
> in the last couple of months.

Fair enough, I couldn't find the discussion where 2.22 was decided to
be the maximum requirement, but SLES11 seems to be the reason then.

What about just using the macro, but not bumping glib requirement?

>
> Regards,
> Daniel
> --
> |: http://berrange.com      -o-    http://www.flickr.com/photos/dberrange/ :|
> |: http://libvirt.org              -o-             http://virt-manager.org :|
> |: http://autobuild.org       -o-         http://search.cpan.org/~danberr/ :|
> |: http://entangle-photo.org       -o-       http://live.gnome.org/gtk-vnc :|
>



-- 
Marc-André Lureau

Re: [Qemu-devel] [PATCH 0/3] Use glib 2.26 version macros

[Daniel P. Berrange]

On Mon, Oct 12, 2015 at 02:24:36PM +0200, Marc-André Lureau wrote:
> Hi
> 
> On Mon, Oct 12, 2015 at 1:26 PM, Daniel P. Berrange <berrange@redhat.com> wrote:
> > On Mon, Oct 12, 2015 at 12:54:56PM +0200, marcandre.lureau@redhat.com wrote:
> >> From: Marc-André Lureau <marcandre.lureau@redhat.com>
> >>
> >> Bump glib requirement to be able to use the GLIB_VERSION macros.
> >
> > We had a discussion about glib min version a few months back where it
> > was decided 2.22 was the maximum we could use at the current time
> > without us loosing support for certain platforms we care about.
> >
> > Simply being able to use the GLIB_VERSION macro does not seem like a
> > compelling enough reason to re-vist this min version again right now,
> > as I doubt we've changed our mind on what platforms we wish to target
> > in the last couple of months.
> 
> Fair enough, I couldn't find the discussion where 2.22 was decided to
> be the maximum requirement, but SLES11 seems to be the reason then.
> 
> What about just using the macro, but not bumping glib requirement?

It would be desirable -  I think someone explored using the GLIB_VERSION
macros before, without bumping version. I can't remember what the problem
was that made them abandon that though - maybe someone else recalls...


Regards,
Daniel
-- 
|: http://berrange.com      -o-    http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org              -o-             http://virt-manager.org :|
|: http://autobuild.org       -o-         http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org       -o-       http://live.gnome.org/gtk-vnc :|

Re: [Qemu-devel] [PATCH 0/3] Use glib 2.26 version macros

[Peter Maydell]

On 12 October 2015 at 13:31, Daniel P. Berrange <berrange@redhat.com> wrote:
> On Mon, Oct 12, 2015 at 02:24:36PM +0200, Marc-André Lureau wrote:
>> What about just using the macro, but not bumping glib requirement?
>
> It would be desirable -  I think someone explored using the GLIB_VERSION
> macros before, without bumping version. I can't remember what the problem
> was that made them abandon that though - maybe someone else recalls...

I think that was David Gilbert, and I vaguely recall there being
bad interactions with our glib-compat.h header.

-- PMM

Re: [Qemu-devel] [PATCH 1/3] hw/acpi/aml-build: remove useless glib version check

[Marc-André Lureau]

to qemu-trivial

On Mon, Oct 12, 2015 at 12:54 PM,  <marcandre.lureau@redhat.com> wrote:
> From: Marc-André Lureau <marcandre.lureau@redhat.com>
>
> 2.22 is the minimum version required
>
> Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
> ---
>  hw/acpi/aml-build.c | 2 --
>  1 file changed, 2 deletions(-)
>
> diff --git a/hw/acpi/aml-build.c b/hw/acpi/aml-build.c
> index 0d4b324..a00a0ab 100644
> --- a/hw/acpi/aml-build.c
> +++ b/hw/acpi/aml-build.c
> @@ -1163,9 +1163,7 @@ void *acpi_data_push(GArray *table_data, unsigned size)
>
>  unsigned acpi_data_len(GArray *table)
>  {
> -#if GLIB_CHECK_VERSION(2, 22, 0)
>      assert(g_array_get_element_size(table) == 1);
> -#endif
>      return table->len;
>  }
>
> --
> 2.4.3
>



-- 
Marc-André Lureau

Re: [Qemu-devel] [PATCH 0/3] Use glib 2.26 version macros

[John Snow]

On 10/12/2015 08:38 AM, Peter Maydell wrote:
> On 12 October 2015 at 13:31, Daniel P. Berrange <berrange@redhat.com> wrote:
>> On Mon, Oct 12, 2015 at 02:24:36PM +0200, Marc-André Lureau wrote:
>>> What about just using the macro, but not bumping glib requirement?
>>
>> It would be desirable -  I think someone explored using the GLIB_VERSION
>> macros before, without bumping version. I can't remember what the problem
>> was that made them abandon that though - maybe someone else recalls...
> 
> I think that was David Gilbert, and I vaguely recall there being
> bad interactions with our glib-compat.h header.
> 
> -- PMM
> 

The version check macros will complain about our back-porting of glib
functions as seen in glib-compat, yes. I think David Gilbert and I each
had a stab at it.

It's fine if you're running 2.22 or so, but if you are running a more
modern version that actually provides the functions we backport/emulate,
the version check macros will catch that and set an error.

Not sure if there's some sort of squelch mechanism. "I know I told you
to complain if I use a 2.23+ function, but forgive me just this once."

--js

Re: [Qemu-devel] [PATCH 0/3] Use glib 2.26 version macros

[Dr. David Alan Gilbert]

* Peter Maydell (peter.maydell@linaro.org) wrote:
> On 12 October 2015 at 13:31, Daniel P. Berrange <berrange@redhat.com> wrote:
> > On Mon, Oct 12, 2015 at 02:24:36PM +0200, Marc-André Lureau wrote:
> >> What about just using the macro, but not bumping glib requirement?
> >
> > It would be desirable -  I think someone explored using the GLIB_VERSION
> > macros before, without bumping version. I can't remember what the problem
> > was that made them abandon that though - maybe someone else recalls...
> 
> I think that was David Gilbert, and I vaguely recall there being
> bad interactions with our glib-compat.h header.

My notes and memory are both somewhat fuzzy on it.

I think that the GLIB_VERSION_MIN/MAX macros only went in in 2.32
but then once you have them they offer a range allowing you to state you
need a minimum of 2.26;  so if we bump the minimum to 2.26
then becareful to check it still works at that.

The interaction of stuff that includes glib.h and others that
include glib-compat.h was messy; I was working on the assumption
that stuff that included glib-compat.h knew it was using newer
things and didn't expect a warning; the problem that Peter
pointed out was that glib-compat gets included by qemu-common so
actually the warnings get disabled in loads of places.

Dave
> 
> -- PMM
--
Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK

Re: [Qemu-devel] [PATCH 0/3] Use glib 2.26 version macros

[Peter Maydell]

On 12 October 2015 at 17:34, Dr. David Alan Gilbert <dgilbert@redhat.com> wrote:
> The interaction of stuff that includes glib.h and others that
> include glib-compat.h was messy; I was working on the assumption
> that stuff that included glib-compat.h knew it was using newer
> things and didn't expect a warning; the problem that Peter
> pointed out was that glib-compat gets included by qemu-common so
> actually the warnings get disabled in loads of places.

Yep; I also want glib-compat to be included by osdep.h and
every .c file to include osdep.h, at which point every source
file ends up including it.

-- PMM

Re: [Qemu-devel] [PATCH 1/3] hw/acpi/aml-build: remove useless glib version check

[Michael Tokarev]

12.10.2015 16:31, Marc-André Lureau wrote:
> to qemu-trivial
> 
> On Mon, Oct 12, 2015 at 12:54 PM,  <marcandre.lureau@redhat.com> wrote:
>> From: Marc-André Lureau <marcandre.lureau@redhat.com>
>>
>> 2.22 is the minimum version required

Applied to -trivial, thank you!

/mjt