Message ID | de75aa7f5d075edcaa2c25e4d021bd5b105e2bb1.1438297123.git.yann.morin.1998@free.fr |
---|---|
State | Rejected |
Headers | show |
Yann, On Fri, 31 Jul 2015 00:59:05 +0200, Yann E. MORIN wrote: > ifeq ($(BR2_PACKAGE_MESA3D_DRI_DRIVER),y) > -XSERVER_XORG_SERVER_CONF_OPTS += --enable-dri --enable-libdrm --enable-glx > -XSERVER_XORG_SERVER_DEPENDENCIES += libdrm mesa3d xproto_xf86driproto > +XSERVER_XORG_SERVER_CONF_OPTS += --enable-dri --enable-glx > +XSERVER_XORG_SERVER_DEPENDENCIES += mesa3d xproto_xf86driproto > +XSERVER_XORG_SERVER_NEEDS_LIBDRM = y Hum, didn't we say that this should be a hidden Config.in boolean? Thomas
Thomas, All, On 2015-07-31 07:57 +0200, Thomas Petazzoni spake thusly: > On Fri, 31 Jul 2015 00:59:05 +0200, Yann E. MORIN wrote: > > ifeq ($(BR2_PACKAGE_MESA3D_DRI_DRIVER),y) > > -XSERVER_XORG_SERVER_CONF_OPTS += --enable-dri --enable-libdrm --enable-glx > > -XSERVER_XORG_SERVER_DEPENDENCIES += libdrm mesa3d xproto_xf86driproto > > +XSERVER_XORG_SERVER_CONF_OPTS += --enable-dri --enable-glx > > +XSERVER_XORG_SERVER_DEPENDENCIES += mesa3d xproto_xf86driproto > > +XSERVER_XORG_SERVER_NEEDS_LIBDRM = y > > Hum, didn't we say that this should be a hidden Config.in boolean? Yes, you suggested so; I said yes; I tried; I said no... ;-) http://lists.busybox.net/pipermail/buildroot/2015-July/135100.html Basically, all the conditional code is in the .mk , and there's almost no conditional in the Config.in. So, I'd like not to duplicate the decision making in Config.in and keep all in a single place, the .mk . That's why I did not move that to a blind kconfig option. Regards, Yann E. MORIN.
Thomas, All, On 2015-07-31 18:59 +0200, Yann E. MORIN spake thusly: > On 2015-07-31 07:57 +0200, Thomas Petazzoni spake thusly: > > On Fri, 31 Jul 2015 00:59:05 +0200, Yann E. MORIN wrote: > > > ifeq ($(BR2_PACKAGE_MESA3D_DRI_DRIVER),y) > > > -XSERVER_XORG_SERVER_CONF_OPTS += --enable-dri --enable-libdrm --enable-glx > > > -XSERVER_XORG_SERVER_DEPENDENCIES += libdrm mesa3d xproto_xf86driproto > > > +XSERVER_XORG_SERVER_CONF_OPTS += --enable-dri --enable-glx > > > +XSERVER_XORG_SERVER_DEPENDENCIES += mesa3d xproto_xf86driproto > > > +XSERVER_XORG_SERVER_NEEDS_LIBDRM = y > > > > Hum, didn't we say that this should be a hidden Config.in boolean? > > Yes, you suggested so; I said yes; I tried; I said no... ;-) > http://lists.busybox.net/pipermail/buildroot/2015-July/135100.html > > Basically, all the conditional code is in the .mk , and there's almost > no conditional in the Config.in. > > So, I'd like not to duplicate the decision making in Config.in and keep > all in a single place, the .mk . > > That's why I did not move that to a blind kconfig option. Did you make up your mind about this? Now that we've switched to a mdular server by default, this series is now fixing actual autobuild errors... If you really prefer that we duplicate all the logic in Config.in , I'll do it. Regards, Yann E. MORIN.
Am Wed, 05 Aug 2015 22:11:05 +0200 schrieb Yann E. MORIN: > Now that we've switched to a mdular server by default, this series is > now fixing actual autobuild errors... Hi Yann, I posted an alternative solution for the build errors: http://patchwork.ozlabs.org/patch/505307/ http://patchwork.ozlabs.org/patch/505305/ http://patchwork.ozlabs.org/patch/505306/ Regards, Bernd
diff --git a/package/x11r7/xserver_xorg-server/xserver_xorg-server.mk b/package/x11r7/xserver_xorg-server/xserver_xorg-server.mk index 79d663c..374a9c3 100644 --- a/package/x11r7/xserver_xorg-server/xserver_xorg-server.mk +++ b/package/x11r7/xserver_xorg-server/xserver_xorg-server.mk @@ -122,10 +122,11 @@ XSERVER_XORG_SERVER_CONF_OPTS += --disable-kdrive --disable-xfbdev endif ifeq ($(BR2_PACKAGE_MESA3D_DRI_DRIVER),y) -XSERVER_XORG_SERVER_CONF_OPTS += --enable-dri --enable-libdrm --enable-glx -XSERVER_XORG_SERVER_DEPENDENCIES += libdrm mesa3d xproto_xf86driproto +XSERVER_XORG_SERVER_CONF_OPTS += --enable-dri --enable-glx +XSERVER_XORG_SERVER_DEPENDENCIES += mesa3d xproto_xf86driproto +XSERVER_XORG_SERVER_NEEDS_LIBDRM = y else -XSERVER_XORG_SERVER_CONF_OPTS += --disable-dri --disable-libdrm --disable-glx +XSERVER_XORG_SERVER_CONF_OPTS += --disable-dri --disable-glx endif ifeq ($(BR2_PACKAGE_XSERVER_XORG_SERVER_AIGLX),y) @@ -145,8 +146,8 @@ XSERVER_XORG_SERVER_DEPENDENCIES += udev XSERVER_XORG_SERVER_CONF_OPTS += --enable-config-udev # udev kms support depends on libdrm ifeq ($(BR2_PACKAGE_LIBDRM),y) -XSERVER_XORG_SERVER_DEPENDENCIES += libdrm -XSERVER_XORG_SERVER_CONF_OPTS += --enable-config-udev-kms --enable-libdrm +XSERVER_XORG_SERVER_CONF_OPTS += --enable-config-udev-kms +XSERVER_XORG_SERVER_NEEDS_LIBDRM = y else XSERVER_XORG_SERVER_CONF_OPTS += --disable-config-udev-kms endif @@ -228,4 +229,11 @@ XSERVER_XORG_SERVER_CONF_OPTS += --with-sha1=libsha1 XSERVER_XORG_SERVER_DEPENDENCIES += libsha1 endif +ifeq ($(XSERVER_XORG_SERVER_NEEDS_LIBDRM),y) +XSERVER_XORG_SERVER_DEPENDENCIES += libdrm +XSERVER_XORG_SERVER_CONF_OPTS += --enable-libdrm +else +XSERVER_XORG_SERVER_CONF_OPTS += --disable-libdrm +endif + $(eval $(autotools-package))
Currently, we have two locations that may want to enable and.or disable libdrm, and it's not clear how those two location correlate one to the other, i.e. they may well want o do one the opposite of the other. Introduce an intermediate variable that records whether libdrm is needed, and use that to depend on, and enable, libdrm. Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr> Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Cc: Bernd Kuhls <bernd.kuhls@t-online.de> --- Note: there might be other libs in the same situation, but that can come in later patches as the need arises... I just change this one lbdrm, because it will be needed by a followup patch to fix building with dri2proto enabled. --- .../x11r7/xserver_xorg-server/xserver_xorg-server.mk | 18 +++++++++++++----- 1 file changed, 13 insertions(+), 5 deletions(-)