[Buildroot] [PATCH 1/1 v2] xdriver_xf86-video-intel: Bump version to 2.99.911

[Buildroot] [PATCH 1/1 v2] xdriver_xf86-video-intel: Bump version to 2.99.911

[Bernd Kuhls]

- add necessary dependencies
- add -D_GNU_SOURCE to avoid getline compilation error

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
---
v2: - depend on mesa3d only when the i965 DRI driver is selected (Thomas)
    - allow both kms and ums support (Thomas):
  Support for Kernel Mode Setting? yes
  Support for legacy User Mode Setting (for i810)? yes

 package/x11r7/xdriver_xf86-video-intel/Config.in   |   13 +++++----
 .../xdriver_xf86-video-intel.mk                    |   29 ++++++++++++++++----
 2 files changed, 30 insertions(+), 12 deletions(-)

diff --git a/package/x11r7/xdriver_xf86-video-intel/Config.in b/package/x11r7/xdriver_xf86-video-intel/Config.in
index 83b714a..4b622df 100644
--- a/package/x11r7/xdriver_xf86-video-intel/Config.in
+++ b/package/x11r7/xdriver_xf86-video-intel/Config.in
@@ -1,13 +1,14 @@
 config BR2_PACKAGE_XDRIVER_XF86_VIDEO_INTEL
 	bool "xf86-video-intel"
-	select BR2_PACKAGE_XPROTO_FONTSPROTO
-	select BR2_PACKAGE_XPROTO_RANDRPROTO
-	select BR2_PACKAGE_XPROTO_RENDERPROTO
-	select BR2_PACKAGE_XPROTO_XPROTO
-	select BR2_PACKAGE_XLIB_LIBPCIACCESS
 	select BR2_PACKAGE_LIBDRM
 	select BR2_PACKAGE_LIBDRM_INTEL
-	select BR2_PACKAGE_MESA3D
+	select BR2_PACKAGE_XLIB_LIBXCOMPOSITE
+	select BR2_PACKAGE_XLIB_LIBPCIACCESS
+	select BR2_PACKAGE_XLIB_LIBXRANDR
+	select BR2_PACKAGE_XLIB_LIBXTST
+	select BR2_PACKAGE_XPROTO_FONTSPROTO
+	select BR2_PACKAGE_XPROTO_XF86DRIPROTO
+	select BR2_PACKAGE_XPROTO_XPROTO
 	depends on (BR2_i386 || BR2_x86_64)
 	help
 	  Intel video driver
diff --git a/package/x11r7/xdriver_xf86-video-intel/xdriver_xf86-video-intel.mk b/package/x11r7/xdriver_xf86-video-intel/xdriver_xf86-video-intel.mk
index 332de54..addb6ec 100644
--- a/package/x11r7/xdriver_xf86-video-intel/xdriver_xf86-video-intel.mk
+++ b/package/x11r7/xdriver_xf86-video-intel/xdriver_xf86-video-intel.mk
@@ -4,18 +4,18 @@
 #
 ################################################################################
 
-XDRIVER_XF86_VIDEO_INTEL_VERSION = 2.99.907
+XDRIVER_XF86_VIDEO_INTEL_VERSION = 2.99.911
 XDRIVER_XF86_VIDEO_INTEL_SOURCE = xf86-video-intel-$(XDRIVER_XF86_VIDEO_INTEL_VERSION).tar.bz2
 XDRIVER_XF86_VIDEO_INTEL_SITE = http://xorg.freedesktop.org/releases/individual/driver
 XDRIVER_XF86_VIDEO_INTEL_LICENSE = MIT
 XDRIVER_XF86_VIDEO_INTEL_LICENSE_FILES = COPYING
-XDRIVER_XF86_VIDEO_INTEL_AUTORECONF = YES
+
+# this fixes a getline-related compilation in src/sna/kgem.c
+XDRIVER_XF86_VIDEO_INTEL_CONF_ENV = \
+	CFLAGS="$(TARGET_CFLAGS) -D_GNU_SOURCE"
 
 XDRIVER_XF86_VIDEO_INTEL_CONF_OPT = \
-	--enable-dri \
-	--enable-kms-only \
 	--disable-xvmc \
-	--disable-ums-only \
 	--enable-sna \
 	--disable-uxa \
 	--disable-glamor \
@@ -23,6 +23,23 @@ XDRIVER_XF86_VIDEO_INTEL_CONF_OPT = \
 	--disable-dga \
 	--disable-async-swap
 
-XDRIVER_XF86_VIDEO_INTEL_DEPENDENCIES = xserver_xorg-server xproto_fontsproto xproto_randrproto xproto_renderproto xproto_xproto libdrm xlib_libpciaccess
+XDRIVER_XF86_VIDEO_INTEL_DEPENDENCIES = \
+	libdrm \
+	xlib_libXcomposite \
+	xlib_libXrandr \
+	xlib_libXtst \
+	xlib_libpciaccess \
+	xproto_fontsproto \
+	xproto_xproto \
+	xserver_xorg-server
+
+ifeq ($(BR2_PACKAGE_MESA3D_DRI_DRIVER_I965),y)
+XDRIVER_XF86_VIDEO_INTEL_CONF_OPT += \
+	--enable-dri
+XDRIVER_XF86_VIDEO_INTEL_DEPENDENCIES += mesa3d
+else
+XDRIVER_XF86_VIDEO_INTEL_CONF_OPT += \
+	--disable-dri
+endif
 
 $(eval $(autotools-package))
-- 
1.7.10.4

[Buildroot] [PATCH 1/1 v2] xdriver_xf86-video-intel: Bump version to 2.99.911

[Thomas Petazzoni]

Dear Bernd Kuhls,

On Sun, 30 Mar 2014 13:22:25 +0200, Bernd Kuhls wrote:

> -XDRIVER_XF86_VIDEO_INTEL_VERSION = 2.99.907
> +XDRIVER_XF86_VIDEO_INTEL_VERSION = 2.99.911
>  XDRIVER_XF86_VIDEO_INTEL_SOURCE = xf86-video-intel-$(XDRIVER_XF86_VIDEO_INTEL_VERSION).tar.bz2
>  XDRIVER_XF86_VIDEO_INTEL_SITE = http://xorg.freedesktop.org/releases/individual/driver
>  XDRIVER_XF86_VIDEO_INTEL_LICENSE = MIT
>  XDRIVER_XF86_VIDEO_INTEL_LICENSE_FILES = COPYING
> -XDRIVER_XF86_VIDEO_INTEL_AUTORECONF = YES
> +
> +# this fixes a getline-related compilation in src/sna/kgem.c
> +XDRIVER_XF86_VIDEO_INTEL_CONF_ENV = \
> +	CFLAGS="$(TARGET_CFLAGS) -D_GNU_SOURCE"
>  
>  XDRIVER_XF86_VIDEO_INTEL_CONF_OPT = \
> -	--enable-dri \
> -	--enable-kms-only \
>  	--disable-xvmc \
> -	--disable-ums-only \

Do we want KMS vs. UMS to be a configurable option?

> +ifeq ($(BR2_PACKAGE_MESA3D_DRI_DRIVER_I965),y)
> +XDRIVER_XF86_VIDEO_INTEL_CONF_OPT += \
> +	--enable-dri
> +XDRIVER_XF86_VIDEO_INTEL_DEPENDENCIES += mesa3d

Then in the Config.in file, if BR2_PACKAGE_MESA3D_DRI_DRIVER_I965 is
enabled, you want to select MESA3D.

Also, in the Config.in file, you select select
BR2_PACKAGE_XPROTO_XF86DRIPROTO, but this package isn't listed in
XDRIVER_XF86_VIDEO_INTEL_DEPENDENCIES as far as I can see.

Best regards,

Thomas
-- 
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com

[Buildroot] [PATCH 1/1 v2] xdriver_xf86-video-intel: Bump version to 2.99.911

[Bernd Kuhls]

Hi,

Thomas Petazzoni
<thomas.petazzoni@free-electrons.com> wrote
in news:20140330151109.658eb835 at skate: 

> Do we want KMS vs. UMS to be a configurable option?

this is not necessary because the driver can have both options compiled, 
therefore I removed the kms/ums-related options.

> Then in the Config.in file, if BR2_PACKAGE_MESA3D_DRI_DRIVER_I965 is
> enabled, you want to select MESA3D.

+       select BR2_PACKAGE_MESA3D if BR2_PACKAGE_MESA3D_DRI_DRIVER_I965

leads to

package/mesa3d/Config.in:1:error: recursive dependency detected!
package/mesa3d/Config.in:1:     symbol BR2_PACKAGE_MESA3D is selected by BR2
_PACKAGE_MESA3D_DRI_DRIVER_I965
package/mesa3d/Config.in:70:    symbol BR2_PACKAGE_MESA3D_DRI_DRIVER_I965 
depends on BR2_PACKAGE_MESA3D

Regards, Bernd

[Buildroot] [PATCH 1/1 v2] xdriver_xf86-video-intel: Bump version to 2.99.911

[Thomas Petazzoni]

Dear Bernd Kuhls,

On Sun, 30 Mar 2014 15:37:33 +0200, Bernd Kuhls wrote:

> > Do we want KMS vs. UMS to be a configurable option?
> 
> this is not necessary because the driver can have both options compiled, 
> therefore I removed the kms/ums-related options.

Ok.

> > Then in the Config.in file, if BR2_PACKAGE_MESA3D_DRI_DRIVER_I965 is
> > enabled, you want to select MESA3D.
> 
> +       select BR2_PACKAGE_MESA3D if BR2_PACKAGE_MESA3D_DRI_DRIVER_I965
> 
> leads to
> 
> package/mesa3d/Config.in:1:error: recursive dependency detected!
> package/mesa3d/Config.in:1:     symbol BR2_PACKAGE_MESA3D is selected by BR2
> _PACKAGE_MESA3D_DRI_DRIVER_I965
> package/mesa3d/Config.in:70:    symbol BR2_PACKAGE_MESA3D_DRI_DRIVER_I965 
> depends on BR2_PACKAGE_MESA3D

Aaah, yes, I was mistaken. If BR2_PACKAGE_MESA3D_DRI_DRIVER_I965 is
enabled, then of course BR2_PACKAGE_MESA3D is already enabled. Sorry
for the noise.

Can you look into the other circular dependency that your patch series
has introduced? I've reported it to you yesterday evening.

Also, by any chance, are you on IRC ? It would sometimes be useful to
be able to discuss things directly with you. Thanks!

Best regards,

Thomas
-- 
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com

[Buildroot] [PATCH 1/1 v2] xdriver_xf86-video-intel: Bump version to 2.99.911

[Bernd Kuhls]

Thomas Petazzoni
<thomas.petazzoni@free-electrons.com> wrote
in news:20140330172413.049cd7d0 at skate: 

> Can you look into the other circular dependency that your patch series
> has introduced? I've reported it to you yesterday evening.

Hi,

I did, but my knowledge about Kconfig and virtual packages unfortunately 
reaches not that far:( One observation though I would like to share: when 
commenting out

source "package/xbmc/Config.in"

the circular dependency error did not show up. Also Yann?s idea resolved the 
"circular hell": 
http://article.gmane.org/gmane.comp.lib.uclibc.buildroot/80652

During the coding on the mesa3d update shortly I already had a virtual 
package for OpenGL, like you proposed, but, as I said, my knowledge here is 
limited so I removed that (non-working) part from the patch in order to have 
a working solution for mesa3d. Since I tested without the xbmc package from 
Maxime (mine is different) I did not notice the dependency problem before.

Regards, Bernd