Buildroot Archive on lore.kernel.org
 help / color / mirror / Atom feed
* [Buildroot] [PATCH] Revert "lvm2: disable build with uClibc-ng"
From: Thomas Petazzoni @ 2017-04-29 12:52 UTC (permalink / raw)
  To: buildroot
In-Reply-To: <20170426072646.GA31873@waldemar-brodkorb.de>

Hello,

On Wed, 26 Apr 2017 09:26:47 +0200, Waldemar Brodkorb wrote:
> This reverts commit 66f136a9f64b880eefa769c1261916773d8f28c0.
> 
> Issue fixed with uClibc-ng update to 1.0.24.
> 
> Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
> ---
>  package/cryptsetup/Config.in    | 5 ++---
>  package/dmraid/Config.in        | 5 ++---
>  package/docker-engine/Config.in | 1 -
>  package/lvm2/Config.in          | 6 ++----
>  package/udisks/Config.in        | 6 ++----
>  5 files changed, 8 insertions(+), 15 deletions(-)

I have started a rebuild of all uClibc toolchains with 1.0.24 + wordexp
enabled. Once those toolchains are ready and deployed, I will apply
this patch. If I apply it before we will have lots of build issues in
the autobuilders.

Best regards,

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

^ permalink raw reply

* [Buildroot] [PATCH 2/2] ne10: Add aarch64 configuration option
From: jerry at chordia.co.uk @ 2017-04-29 12:43 UTC (permalink / raw)
  To: buildroot
In-Reply-To: <20170429140953.0b7badef@free-electrons.com>

Hello,

> 
> But I believe you should in fact rework your patches:
> 
>  - One patch that just bumps ne10, keeping only the ARM support
> 
>  - One patch that refactors the architecture dependency with a
>    Config.in hidden boolean BR2_PACKAGE_NE10_ARCH_SUPPORTS
> 
>  - One patch that allows to use ne10 on AArch64

OK, I will do that.

BR

Jerry

^ permalink raw reply

* [Buildroot] [PATCH 1/1] package/kodi: optimise libva/libvdpau dependencies
From: Thomas Petazzoni @ 2017-04-29 12:43 UTC (permalink / raw)
  To: buildroot
In-Reply-To: <20170429122732.29011-1-bernd.kuhls@t-online.de>

Hello,

On Sat, 29 Apr 2017 14:27:32 +0200, Bernd Kuhls wrote:
> Suggested by Thomas:
> http://lists.busybox.net/pipermail/buildroot/2017-April/190703.html
> 
> Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
> ---
>  package/kodi/Config.in | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)

Applied to master, thanks.

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

^ permalink raw reply

* [Buildroot] [git commit] package/kodi: optimise libva/libvdpau dependencies
From: Thomas Petazzoni @ 2017-04-29 12:42 UTC (permalink / raw)
  To: buildroot

commit: https://git.buildroot.net/buildroot/commit/?id=b556a9e06a05af9a064de138dbba7e109e1cd6db
branch: https://git.buildroot.net/buildroot/commit/?id=refs/heads/master

Suggested by Thomas:
http://lists.busybox.net/pipermail/buildroot/2017-April/190703.html

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
---
 package/kodi/Config.in | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/package/kodi/Config.in b/package/kodi/Config.in
index fc40343..613cf48 100644
--- a/package/kodi/Config.in
+++ b/package/kodi/Config.in
@@ -241,25 +241,25 @@ config BR2_PACKAGE_KODI_LIBUSB
 
 config BR2_PACKAGE_KODI_LIBVA
 	bool "va"
-	depends on !BR2_PACKAGE_KODI_EGL_GLES
+	depends on BR2_PACKAGE_KODI_EGL_GL
 	depends on BR2_PACKAGE_XORG7
 	select BR2_PACKAGE_LIBVA
 	help
 	  Enable libva support.
 
 comment "libva support needs X.org with an openGL backend"
-	depends on !BR2_PACKAGE_XORG7 || BR2_PACKAGE_KODI_EGL_GLES
+	depends on !BR2_PACKAGE_XORG7 || !BR2_PACKAGE_KODI_EGL_GL
 
 config BR2_PACKAGE_KODI_LIBVDPAU
 	bool "vdpau"
-	depends on !BR2_PACKAGE_KODI_EGL_GLES
+	depends on BR2_PACKAGE_KODI_EGL_GL
 	depends on BR2_PACKAGE_XORG7
 	select BR2_PACKAGE_LIBVDPAU
 	help
 	  Enable libvdpau support.
 
 comment "libvdpau support needs X.org with an openGL backend"
-	depends on !BR2_PACKAGE_XORG7 || BR2_PACKAGE_KODI_EGL_GLES
+	depends on !BR2_PACKAGE_XORG7 || !BR2_PACKAGE_KODI_EGL_GL
 
 config BR2_PACKAGE_KODI_UPNP
 	bool "upnp"

^ permalink raw reply related

* [Buildroot] [PATCH v7 08/31] package/kodi: bump to version 17.1-Krypton
From: Thomas Petazzoni @ 2017-04-29 12:42 UTC (permalink / raw)
  To: buildroot
In-Reply-To: <fe62a0d1-d466-2d4d-22c4-d5165882dc7b@t-online.de>

Hello,

On Sat, 29 Apr 2017 14:29:13 +0200, Bernd Kuhls wrote:

> > I don't understand why you have this !BR2_PACKAGE_RPI_USERLAND
> > dependency here. Can you explain?  
> 
> We need to rule out that a rbpi build is done with opengl/egl, otherwise
> the build will fail like this:
> http://forum.kodi.tv/showthread.php?tid=305752&pid=2569701#pid2569701

But rpi-userland doesn't provide OpenGL support, only OpenGLES, so how
is it possible to get a RPi build with OpenGL support ?


> Ok, sent patch: http://patchwork.ozlabs.org/patch/756733/

Thanks, looks good.


> > This looks a bit fishy. Why each option has a _OK option and a _TRUE
> > option? Is setting both really needed?  
> 
> Yes, otherwise host CPU features sneak into the build parameters,
> upstream does something similar:
> https://github.com/xbmc/xbmc/blob/Krypton/project/cmake/modules/
> FindSSE.cmake#L81

Hum, OK.

> > This looks very very weird. Why is LDGOLD related to RPi ?  
> 
> RPi and Linux are considered different platforms, both have their own
> ArchSetup.cmake:
> 
> https://github.com/xbmc/xbmc/blob/Krypton/project/cmake/scripts/linux/
> ArchSetup.cmake
> https://github.com/xbmc/xbmc/blob/Krypton/project/cmake/scripts/rbpi/
> ArchSetup.cmake
> 
> LDGOLD is only used for the linux platform:
> https://github.com/xbmc/xbmc/blob/Krypton/project/cmake/scripts/linux/
> ArchSetup.cmake#L34
> 
> If you think that treating Linux and RPi as separate platforms is weird
> then you are not alone, but it may change in Kodi 18:
> https://github.com/xbmc/xbmc/pull/12021 ;)

Ah, ok, makes sense. Well, considering them as separate platforms
doesn't, but the way you've handled the current situation makes sense.
What about adding a comment?

Also, I changed rbpi to rpi in the comments, thinking that rbpi was a
typo, but it apparently isn't.

> Because all these options do not exist in https://github.com/xbmc/xbmc/
> blob/Krypton/project/cmake/scripts/rbpi/ArchSetup.cmake

OK. Good reason to add a comment in the .mk file.

> > So this thing is the only reason why we need xmlstarlet built for the
> > host system?  
> 
> No, I plan to add more patches to control the skin addon(s) being used,
> also the default skin addon can then be changed using xmlstarlet:
> http://lists.busybox.net/pipermail/buildroot/2017-April/189753.html
> 
> I removed these patches from v6+ because Yann had problems with them:
> http://lists.busybox.net/pipermail/buildroot/2017-April/190190.html
> and, as feature patches, they should not block the version bump.

It is indeed a good idea to have separate those changes from the
version bump itself. Now the version bump has been merged, it is
possible to focus on this rather than having a too long patch series.

Thanks!

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

^ permalink raw reply

* [Buildroot] [PATCH 1/2] x11r7: xdriver_xf86-input-tslib: new upstream location
From: Martin Kepplinger @ 2017-04-29 12:42 UTC (permalink / raw)
  To: buildroot
In-Reply-To: <20170429135844.2fa76511@free-electrons.com>

Am 2017-04-29 um 13:58 schrieb Thomas Petazzoni:
> Hello,
> 
> On Fri, 28 Apr 2017 14:36:15 +0200, Martin Kepplinger wrote:
> 
>>  XDRIVER_XF86_INPUT_TSLIB_VERSION = 0.0.6
>>  XDRIVER_XF86_INPUT_TSLIB_SOURCE = xf86-input-tslib-$(XDRIVER_XF86_INPUT_TSLIB_VERSION).tar.bz2
>> -XDRIVER_XF86_INPUT_TSLIB_SITE = http://www.pengutronix.de/software/xf86-input-tslib/download
>> +XDRIVER_XF86_INPUT_TSLIB_SITE = https://github.com/merge/xf86-input-tslib/releases/download/$(TSLIB_VERSION)
> 
> This couldn't work: TSLIB_VERSION comes from tslib.mk and is set to
> 1.9, which prevents the download from working. Changing this to
> XDRIVER_XF86_INPUT_TSLIB_VERSION fixed the problem.
> 
> Applied with this fix. Thanks!
> 

oops. sorry. Should have taken more time to check everything. Thanks!

^ permalink raw reply

* [Buildroot] [PATCH] package/protobuf: depend on gcc >= 4.5
From: Thomas Petazzoni @ 2017-04-29 12:31 UTC (permalink / raw)
  To: buildroot
In-Reply-To: <20170426162850.5214-1-mrugiero@gmail.com>

Hello,

On Wed, 26 Apr 2017 13:28:50 -0300, mrugiero at gmail.com wrote:
> From: "Mario J. Rugiero" <mrugiero@gmail.com>
> 
> Protobuf hits a bug in gcc < 4.5 which breaks the build.
> See https://gcc.gnu.org/bugzilla/show_bug.cgi?id=189

This bug is marked as FIXED, and it has been fixed since many years.

What is the actual problem you are seeing? Is there a reference to an
autobuilder issue?

I've tried to build with gcc 4.3, and indeed I'm getting:

./google/protobuf/metadata.h: In constructor 'google::protobuf::internal::InternalMetadataWithArena::InternalMetadataWithArena(google::protobuf::Arena*)':
./google/protobuf/metadata.h:175: error: class 'google::protobuf::internal::InternalMetadataWithArena' does not have any field named 'InternalMetadataWithArenaBase'
./google/protobuf/metadata.h: In constructor 'google::protobuf::internal::InternalMetadataWithArenaLite::InternalMetadataWithArenaLite(google::protobuf::Arena*)':
./google/protobuf/metadata.h:204: error: class 'google::protobuf::internal::InternalMetadataWithArenaLite' does not have any field named 'InternalMetadataWithArenaBase'

Is this the issue we're talking about ?

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

^ permalink raw reply

* [Buildroot] [PATCH v7 08/31] package/kodi: bump to version 17.1-Krypton
From: Bernd Kuhls @ 2017-04-29 12:29 UTC (permalink / raw)
  To: buildroot
In-Reply-To: <20170429115608.0b9bf581@free-electrons.com>

Hi Thomas,

thanks for the commit!

Am Sat, 29 Apr 2017 11:56:08 +0200 schrieb Thomas Petazzoni:

> Hello,
> 
> I've applied, after doing a number of minor tweaks. Also, there are a
> few things that don't look completely good, and that I believe should be
> addressed as follow-up patches.
> 
> On Sat, 29 Apr 2017 10:37:28 +0200, Bernd Kuhls wrote:
> 
>> +config BR2_PACKAGE_KODI_RTMPDUMP
>> +	bool "kodi rtmp"
> 
> This prompt was quite useless, so I've changed it to "kodi rtmp support
> removed".

Ok.

>> -comment "kodi needs an OpenGL or an openGL ES and EGL backend"
>> -	depends on BR2_i386 || BR2_x86_64
>> -	depends on !BR2_PACKAGE_KODI_GL && !BR2_PACKAGE_KODI_EGL_GLES
>> +	depends on !BR2_PACKAGE_RPI_USERLAND # rpi depends on gles
> 
> I don't understand why you have this !BR2_PACKAGE_RPI_USERLAND
> dependency here. Can you explain?

We need to rule out that a rbpi build is done with opengl/egl, otherwise 
the build will fail like this:
http://forum.kodi.tv/showthread.php?tid=305752&pid=2569701#pid2569701

>>  config BR2_PACKAGE_KODI_LIBVA
>>  	bool "va"
>> +	depends on !BR2_PACKAGE_KODI_EGL_GLES
> 
> What about using
> 
> 	depends on BR2_PACKAGE_KODI_EGL_GL
> 
> here ?
> 
>> +	depends on BR2_PACKAGE_XORG7
>>  	select BR2_PACKAGE_LIBVA
>>  	help
>>  	  Enable libva support.
>>  
>> +comment "libva support needs X.org with an openGL backend"
>> +	depends on !BR2_PACKAGE_XORG7 || BR2_PACKAGE_KODI_EGL_GLES
> 
> And ... || !BR2_PACKAGE_KODI_EGL_GL here ?
> 
>>  config BR2_PACKAGE_KODI_LIBVDPAU
>>  	bool "vdpau"
>> +	depends on !BR2_PACKAGE_KODI_EGL_GLES
> 
> Same suggestion.

Ok, sent patch: http://patchwork.ozlabs.org/patch/756733/

>> +KODI_SUBDIR = project/cmake
>> +
>> +KODI_LIBDVDCSS_VERSION = 2f12236
>> +KODI_LIBDVDNAV_VERSION = 981488f
>> +KODI_LIBDVDREAD_VERSION = 17d99db
>> +
>> +KODI_EXTRA_DOWNLOADS = \
>> +	https://github.com/xbmc/libdvdcss/archive/
$(KODI_LIBDVDCSS_VERSION).tar.gz \
>> +	https://github.com/xbmc/libdvdnav/archive/
$(KODI_LIBDVDNAV_VERSION).tar.gz \
>> +	https://github.com/xbmc/libdvdread/archive/
$(KODI_LIBDVDREAD_VERSION).tar.gz
> 
> Can I say again how much I hate this stuff?
> 
>> +KODI_CONF_OPTS += \
>> +	-DLIBDVDCSS_URL=$(BR2_DL_DIR)/$(KODI_LIBDVDCSS_VERSION).tar.gz \
>> +	-DLIBDVDNAV_URL=$(BR2_DL_DIR)/$(KODI_LIBDVDNAV_VERSION).tar.gz \
>> +	-DLIBDVDREAD_URL=$(BR2_DL_DIR)/$(KODI_LIBDVDREAD_VERSION).tar.gz
> 
> I've moved these below, with the other unconditional CONF_OPTS.

Ok.

> 
>> +ifeq ($(BR2_X86_CPU_HAS_SSE),y)
>> +KODI_CONF_OPTS += -D_SSE_OK=ON -D_SSE_TRUE=ON
> 
> This looks a bit fishy. Why each option has a _OK option and a _TRUE
> option? Is setting both really needed?

Yes, otherwise host CPU features sneak into the build parameters,
upstream does something similar:
https://github.com/xbmc/xbmc/blob/Krypton/project/cmake/modules/
FindSSE.cmake#L81

>>  ifeq ($(BR2_PACKAGE_RPI_USERLAND),y)
>> +KODI_CONF_OPTS += -DCORE_SYSTEM_NAME=rbpi
>>  KODI_DEPENDENCIES += rpi-userland
>> -KODI_CONF_OPTS += --with-platform=raspberry-pi --enable-
player=omxplayer
>> -KODI_INCLUDES += \
>> -	-I$(STAGING_DIR)/usr/include/interface/vcos/pthreads \
>> -	-I$(STAGING_DIR)/usr/include/interface/vmcs_host/linux
>> -KODI_LIBS = -lvcos -lvchostif
>> -endif
>> -
>> -ifeq ($(BR2_PACKAGE_HAS_UDEV),y)
>> -KODI_DEPENDENCIES += udev
>> -KODI_CONF_OPTS += --enable-udev
>>  else
>> -KODI_CONF_OPTS += --disable-udev
>> +# these options only exist on non-rbpi systems
> 
> This looks very very weird. Why is LDGOLD related to RPi ?

RPi and Linux are considered different platforms, both have their own 
ArchSetup.cmake:

https://github.com/xbmc/xbmc/blob/Krypton/project/cmake/scripts/linux/
ArchSetup.cmake
https://github.com/xbmc/xbmc/blob/Krypton/project/cmake/scripts/rbpi/
ArchSetup.cmake

LDGOLD is only used for the linux platform:
https://github.com/xbmc/xbmc/blob/Krypton/project/cmake/scripts/linux/
ArchSetup.cmake#L34

If you think that treating Linux and RPi as separate platforms is weird 
then you are not alone, but it may change in Kodi 18:
https://github.com/xbmc/xbmc/pull/12021 ;)

>> +KODI_CONF_OPTS += -DENABLE_LDGOLD=OFF
>> +ifeq ($(BR2_PACKAGE_LIBAMCODEC),y)
> 
> Why should these options be under a "else" of the RPi stuff ?

Because all these options do not exist in https://github.com/xbmc/xbmc/
blob/Krypton/project/cmake/scripts/rbpi/ArchSetup.cmake

>> -ifeq ($(BR2_PACKAGE_KODI_GL),y)
>> -KODI_DEPENDENCIES += libglew libglu libgl xlib_libX11 xlib_libXext \
>> -	xlib_libXmu xlib_libXrandr xlib_libXt libdrm
>> -KODI_CONF_OPTS += --enable-gl --enable-x11 --disable-gles
>> +ifeq ($(BR2_PACKAGE_KODI_GL_EGL),y)
>> +KODI_DEPENDENCIES += libegl libglu libgl xlib_libX11 xlib_libXext \
>> +	xlib_libXrandr libdrm
>> +KODI_CONF_OPTS += -DENABLE_OPENGL=ON -DENABLE_X11=ON -
DENABLE_OPENGLES=OFF
>>  else
>> -KODI_CONF_OPTS += --disable-gl --disable-x11
>> +KODI_CONF_OPTS += -DENABLE_OPENGL=OFF -DENABLE_X11=OFF
>>  ifeq ($(BR2_PACKAGE_KODI_EGL_GLES),y)
> 
> No reason for this being in the "else" of the ifeq
> ($(BR2_PACKAGE_KODI_GL_EGL), so I've re-organized this way:
> 
> ifeq ( ... opengl + egl ...)
>  enable Opengl stuff
> else
>  disable Opengl stuff
> endif
> 
> ifeq ( ... opengles + egl ...)
>  enable OpenglES stuff
> else
>  disable OpenglES stuff
> endif

Ok.

>> +KODI_CONF_OPTS += -DCMAKE_CXX_FLAGS="$(TARGET_CXXFLAGS) `
$(PKG_CONFIG_HOST_BINARY) --cflags --libs egl`"
>> +KODI_CONF_OPTS += -DCMAKE_C_FLAGS="$(TARGET_CFLAGS) `
$(PKG_CONFIG_HOST_BINARY) --cflags --libs egl`"
>> +KODI_CONF_OPTS += -DENABLE_OPENGLES=ON
> 
> Reorganized as one assignment to KODI_CONF_OPTS, i.e:
> 
> KODI_CONF_OPTS += \
> 	... \
> 	... \
> 	...

Ok.

>> +	$(HOST_DIR)/usr/bin/xml ed -L \
>> +		-d "/addons/addon[text()='service.xbmc.versioncheck']" \
>> +		$(KODI_ADDON_MANIFEST)
> 
> So this thing is the only reason why we need xmlstarlet built for the
> host system?

No, I plan to add more patches to control the skin addon(s) being used, 
also the default skin addon can then be changed using xmlstarlet:
http://lists.busybox.net/pipermail/buildroot/2017-April/189753.html

I removed these patches from v6+ because Yann had problems with them:
http://lists.busybox.net/pipermail/buildroot/2017-April/190190.html
and, as feature patches, they should not block the version bump.

Regards, Bernd

^ permalink raw reply

* [Buildroot] [PATCH 1/1] package/kodi: optimise libva/libvdpau dependencies
From: Bernd Kuhls @ 2017-04-29 12:27 UTC (permalink / raw)
  To: buildroot

Suggested by Thomas:
http://lists.busybox.net/pipermail/buildroot/2017-April/190703.html

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
---
 package/kodi/Config.in | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/package/kodi/Config.in b/package/kodi/Config.in
index bf4ea01549..b9458dae68 100644
--- a/package/kodi/Config.in
+++ b/package/kodi/Config.in
@@ -230,25 +230,25 @@ config BR2_PACKAGE_KODI_LIBUSB
 
 config BR2_PACKAGE_KODI_LIBVA
 	bool "va"
-	depends on !BR2_PACKAGE_KODI_EGL_GLES
+	depends on BR2_PACKAGE_KODI_EGL_GL
 	depends on BR2_PACKAGE_XORG7
 	select BR2_PACKAGE_LIBVA
 	help
 	  Enable libva support.
 
 comment "libva support needs X.org with an openGL backend"
-	depends on !BR2_PACKAGE_XORG7 || BR2_PACKAGE_KODI_EGL_GLES
+	depends on !BR2_PACKAGE_XORG7 || !BR2_PACKAGE_KODI_EGL_GL
 
 config BR2_PACKAGE_KODI_LIBVDPAU
 	bool "vdpau"
-	depends on !BR2_PACKAGE_KODI_EGL_GLES
+	depends on BR2_PACKAGE_KODI_EGL_GL
 	depends on BR2_PACKAGE_XORG7
 	select BR2_PACKAGE_LIBVDPAU
 	help
 	  Enable libvdpau support.
 
 comment "libvdpau support needs X.org with an openGL backend"
-	depends on !BR2_PACKAGE_XORG7 || BR2_PACKAGE_KODI_EGL_GLES
+	depends on !BR2_PACKAGE_XORG7 || !BR2_PACKAGE_KODI_EGL_GL
 
 config BR2_PACKAGE_KODI_UPNP
 	bool "upnp"
-- 
2.11.0

^ permalink raw reply related

* [Buildroot] [PATCH v2] package/uclibc: Enable wordexp functionality
From: Thomas Petazzoni @ 2017-04-29 12:26 UTC (permalink / raw)
  To: buildroot
In-Reply-To: <20170321160237.31058-2-andriy.shevchenko@linux.intel.com>

Hello,

On Tue, 21 Mar 2017 18:02:36 +0200, Andy Shevchenko wrote:
> The commit d7c54cb0c0cf ("bluez5_utils: not available with uClibc")
> disables bluez5_utils (most of them) to be built with uClibc. Looking to
> the development of BlueZ user space support that commit basically makes
> the bluez5_utils package somewhat piece of junk. Since bluez5_utils 5.44
> deprecates all previously used tools and focuses on btmgmt and
> bluetoothctl we need to restore status quo.
> 
> Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> ---
>  package/uclibc/uClibc-ng.config | 1 +
>  1 file changed, 1 insertion(+)

I've applied, after rewriting the commit message. If there's some
bluez5_utils improvements that can be done now that wordexp is enabled,
please send patches. Thanks!

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

^ permalink raw reply

* [Buildroot] [git commit] package/uclibc: enable wordexp functionality
From: Thomas Petazzoni @ 2017-04-29 12:25 UTC (permalink / raw)
  To: buildroot

commit: https://git.buildroot.net/buildroot/commit/?id=85b8a8ab9811144f0d65df9f064235c60ea15dee
branch: https://git.buildroot.net/buildroot/commit/?id=refs/heads/master

Wordexp support is needed by more and more packages, recently
bluez5_utils. It adds only ~16 KB to uClibc, so let's add it by default
to keep things simple.

Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Waldemar Brodkorb <wbx@openadk.org>
[Thomas: rework commit message.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
---
 package/uclibc/uClibc-ng.config | 1 +
 1 file changed, 1 insertion(+)

diff --git a/package/uclibc/uClibc-ng.config b/package/uclibc/uClibc-ng.config
index 6780ad6..f0a876f 100644
--- a/package/uclibc/uClibc-ng.config
+++ b/package/uclibc/uClibc-ng.config
@@ -30,6 +30,7 @@ UCLIBC_HAS_STDIO_PUTC_MACRO=y
 UCLIBC_HAS_FOPEN_EXCLUSIVE_MODE=y
 UCLIBC_HAS_GLIBC_CUSTOM_STREAMS=y
 UCLIBC_HAS_PRINTF_M_SPEC=y
+UCLIBC_HAS_WORDEXP=y
 UCLIBC_HAS_NFTW=y
 UCLIBC_HAS_FTW=y
 UCLIBC_HAS_GNU_GLOB=y

^ permalink raw reply related

* [Buildroot] [PATCH 2/2] ne10: Add aarch64 configuration option
From: Thomas Petazzoni @ 2017-04-29 12:09 UTC (permalink / raw)
  To: buildroot
In-Reply-To: <1493418486-109149-2-git-send-email-g4@novadsp.com>

Hello,

On Fri, 28 Apr 2017 23:28:06 +0100, J Evans wrote:

> +ifeq ($(BR2_aarch64),y)
> +NE10_CONF_OPTS += \
> +    -DNE10_LINUX_TARGET_ARCH=aarch64
> +endif

PATCH 1/2 is already enabling ne10 on AArch64, so there's something
wrong going on here. This change should go with the patch enabling ne10
on AArch64.

But I believe you should in fact rework your patches:

 - One patch that just bumps ne10, keeping only the ARM support

 - One patch that refactors the architecture dependency with a
   Config.in hidden boolean BR2_PACKAGE_NE10_ARCH_SUPPORTS

 - One patch that allows to use ne10 on AArch64

Best regards,

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

^ permalink raw reply

* [Buildroot] [PATCH 1/2] ne10: bump to version 1.2.1. Remove obsolete patch file for modifying thumb code generation.
From: Thomas Petazzoni @ 2017-04-29 12:08 UTC (permalink / raw)
  To: buildroot
In-Reply-To: <1493418486-109149-1-git-send-email-g4@novadsp.com>

Hello,

Since the change is not trivial, it requires a more verbose commit log.
You're doing much more than bumping.

Also, the commit title should be shorter, just "ne10: bump to version
1.2.1".

On Fri, 28 Apr 2017 23:28:05 +0100, J Evans wrote:

>  ...sts-don-t-hard-code-thumb-code-generation.patch | 29 ----------------------

What makes you think this patch is no longer needed? The CMakeLists.txt
is still hardcoding -mthumb:

    if("${NE10_TARGET_ARCH}" STREQUAL "armv7")
      set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -mthumb-interwork -mthumb -march=armv7-a -mfpu=vfp3 -funsafe-math-optimizations")
      set(CMAKE_ASM_FLAGS "${CMAKE_C_FLAGS} -mthumb-interwork -mthumb -march=armv7-a -mfpu=neon")


> diff --git a/package/ne10/Config.in b/package/ne10/Config.in
> old mode 100644
> new mode 100755

Why are you making this file executable? Are you working on Windows?

> index 8d37579..ebd23a6
> --- a/package/ne10/Config.in
> +++ b/package/ne10/Config.in
> @@ -1,7 +1,6 @@
>  config BR2_PACKAGE_NE10
>  	bool "ne10"
> -	depends on BR2_arm
> -	depends on BR2_ARM_FPU_NEON || BR2_ARM_FPU_NEON_VFPV4
> +	depends on BR2_aarch64 || (BR2_arm && (BR2_ARM_FPU_NEON || BR2_ARM_FPU_NEON_VFPV4))

I think it would be nicer to introduce a blind option:

config BR2_PACKAGE_NE10_ARCH_SUPPORTS
	bool
	default y if BR2_aarch64
	default y if BR2_arm && (BR2_ARM_FPU_NEON || BR2_ARM_FPU_NEON_VFPV4)

and then use it in BR2_PACKAGE_NE10 and the Config.in comment.


> old mode 100644
> new mode 100755
> index db5b6f1..923e660
> --- a/package/ne10/ne10.hash
> +++ b/package/ne10/ne10.hash

Why are you making this file executable ?

> @@ -1,2 +1,2 @@
> -# Locally calculated
> -sha256 450dac5bb3a2351019ae23792e97c3bf965d16a21c0598b73cea9fbc3b3a56af  ne10-v1.2.0.tar.gz
> +# Locally calculated after checking pgp signature
> +sha256	dd1a96610c0692cc80154ae123edd5d25e6e0a3f81d0c16a96425f3ef57b9929	ne10-v1.2.1.tar.gz
> diff --git a/package/ne10/ne10.mk b/package/ne10/ne10.mk
> old mode 100644
> new mode 100755
> index cb303bd..0da862c
> --- a/package/ne10/ne10.mk
> +++ b/package/ne10/ne10.mk

Same question.

Best regards,

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

^ permalink raw reply

* [Buildroot] [PATCH] ffmpeg: fix coldfire compile issue
From: Thomas Petazzoni @ 2017-04-29 12:03 UTC (permalink / raw)
  To: buildroot
In-Reply-To: <20170429073019.GA22285@waldemar-brodkorb.de>

Hello,

On Sat, 29 Apr 2017 09:30:19 +0200, Waldemar Brodkorb wrote:
> Fixes:
>   http://autobuild.buildroot.net/results/7e95ac6d78bc1732127bea084b4f791e52637abd
> 
> Signed-off-by: Waldemar Brodkorb <wbx@openadk.org>
> ---
>  ...llow-to-directly-use-pthreads-atomics-com.patch | 65 ++++++++++++++++++++++
>  package/ffmpeg/ffmpeg.mk                           |  6 ++
>  2 files changed, 71 insertions(+)
>  create mode 100644 package/ffmpeg/0002-configure-allow-to-directly-use-pthreads-atomics-com.patch

I'm hesitating: do we apply this fix, or simply adjust
BR2_PACKAGE_FFMPEG_ARCH_SUPPORTS do disallow using ffmpeg on
BR2_m68k_cf ? After all, who is realistically going to use ffmpeg on
m68k_cf ?

Best regards,

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

^ permalink raw reply

* [Buildroot] [PATCH 2/2] ne10: Add aarch64 configuration option
From: g4 at novadsp.com @ 2017-04-29 12:00 UTC (permalink / raw)
  To: buildroot
In-Reply-To: <1493418486-109149-2-git-send-email-g4@novadsp.com>

> -----Original Message-----
> From: buildroot [mailto:buildroot-bounces at busybox.net] On Behalf Of J
> Evans
> Sent: 28 April 2017 23:28
> To: buildroot at buildroot.org
> Cc: Thomas Petazzoni; J Evans
> Subject: [Buildroot] [PATCH 2/2] ne10: Add aarch64 configuration option
> 
> Signed-off-by: J Evans <g4@novadsp.com>

Apologies. This patch appears to be incomplete, please ignore. Whilst it
works fine for ARM64 the latest does not want to configure correctly for a
PI3 for example.

BR,

Jerry.

^ permalink raw reply

* [Buildroot] [PATCH] ola: mark as broken
From: Thomas Petazzoni @ 2017-04-29 11:58 UTC (permalink / raw)
  To: buildroot
In-Reply-To: <78a929f7208b78d17f45d4ddb8bc8df7f26ba74b.1493388478.git.baruch@tkos.co.il>

Hello,

On Fri, 28 Apr 2017 17:07:58 +0300, Baruch Siach wrote:
> Build with current protobuf is broken. Mark as broken until upstream resolves
> this issue.
> 
> https://github.com/OpenLightingProject/ola/issues/1192
> 
> Fixes:
> http://autobuild.buildroot.net/results/d9a/d9a24f7b715100be1580a568a5e3ff72b0389165/
> http://autobuild.buildroot.net/results/b31/b314811dedce04ebdc779df67de6cb59a1880cac/
> http://autobuild.buildroot.net/results/587/5877b2301b7da43c50127a4c5f648acd3b0264cc/
> 
> Cc: Dave Skok <blanco.ether@gmail.com>
> Signed-off-by: Baruch Siach <baruch@tkos.co.il>
> ---
>  package/ola/Config.in | 4 ++++
>  1 file changed, 4 insertions(+)

Applied to master, thanks.

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

^ permalink raw reply

* [Buildroot] [PATCH v2] x11r7: xdriver_xf86-input-tslib: update to 0.0.7
From: Thomas Petazzoni @ 2017-04-29 11:58 UTC (permalink / raw)
  To: buildroot
In-Reply-To: <1493386434-26756-1-git-send-email-martink@posteo.de>

Hello,

On Fri, 28 Apr 2017 15:33:54 +0200, Martin Kepplinger wrote:
> The patches are removed as they are part of this release. 0.0.7 is a
> bugfix and compatibility release to keep this usable for on newer systems.
> 
> Signed-off-by: Martin Kepplinger <martink@posteo.de>
> ---

Applied to master, thanks.

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

^ permalink raw reply

* [Buildroot] [PATCH 1/2] x11r7: xdriver_xf86-input-tslib: new upstream location
From: Thomas Petazzoni @ 2017-04-29 11:58 UTC (permalink / raw)
  To: buildroot
In-Reply-To: <1493382976-26007-1-git-send-email-martink@posteo.de>

Hello,

On Fri, 28 Apr 2017 14:36:15 +0200, Martin Kepplinger wrote:

>  XDRIVER_XF86_INPUT_TSLIB_VERSION = 0.0.6
>  XDRIVER_XF86_INPUT_TSLIB_SOURCE = xf86-input-tslib-$(XDRIVER_XF86_INPUT_TSLIB_VERSION).tar.bz2
> -XDRIVER_XF86_INPUT_TSLIB_SITE = http://www.pengutronix.de/software/xf86-input-tslib/download
> +XDRIVER_XF86_INPUT_TSLIB_SITE = https://github.com/merge/xf86-input-tslib/releases/download/$(TSLIB_VERSION)

This couldn't work: TSLIB_VERSION comes from tslib.mk and is set to
1.9, which prevents the download from working. Changing this to
XDRIVER_XF86_INPUT_TSLIB_VERSION fixed the problem.

Applied with this fix. Thanks!

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

^ permalink raw reply

* [Buildroot] [git commit] ola: mark as broken
From: Thomas Petazzoni @ 2017-04-29 11:57 UTC (permalink / raw)
  To: buildroot

commit: https://git.buildroot.net/buildroot/commit/?id=f80fce90b73ed28fbbc1315c3d9a5577523fddf8
branch: https://git.buildroot.net/buildroot/commit/?id=refs/heads/master

Build with current protobuf is broken. Mark as broken until upstream resolves
this issue.

https://github.com/OpenLightingProject/ola/issues/1192

Fixes:
http://autobuild.buildroot.net/results/d9a/d9a24f7b715100be1580a568a5e3ff72b0389165/
http://autobuild.buildroot.net/results/b31/b314811dedce04ebdc779df67de6cb59a1880cac/
http://autobuild.buildroot.net/results/587/5877b2301b7da43c50127a4c5f648acd3b0264cc/

Cc: Dave Skok <blanco.ether@gmail.com>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
---
 package/ola/Config.in | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/package/ola/Config.in b/package/ola/Config.in
index 90a92fd..656e825 100644
--- a/package/ola/Config.in
+++ b/package/ola/Config.in
@@ -1,6 +1,7 @@
 comment "ola needs a toolchain w/ C++, threads, dynamic library"
 	depends on !BR2_INSTALL_LIBSTDCPP || !BR2_TOOLCHAIN_HAS_THREADS \
 		|| BR2_STATIC_LIBS
+	depends on BROKEN
 	depends on BR2_PACKAGE_PROTOBUF_ARCH_SUPPORTS
 
 menuconfig BR2_PACKAGE_OLA
@@ -8,6 +9,9 @@ menuconfig BR2_PACKAGE_OLA
 	select BR2_PACKAGE_PROTOBUF
 	select BR2_PACKAGE_UTIL_LINUX_LIBUUID
 	select BR2_PACKAGE_UTIL_LINUX
+	# Build fails with current protobuf
+	# https://github.com/OpenLightingProject/ola/issues/1192
+	depends on BROKEN
 	depends on BR2_INSTALL_LIBSTDCPP # protobuf
 	depends on !BR2_STATIC_LIBS # protobuf
 	depends on BR2_TOOLCHAIN_HAS_THREADS

^ permalink raw reply related

* [Buildroot] [git commit] x11r7: xdriver_xf86-input-tslib: update to 0.0.7
From: Thomas Petazzoni @ 2017-04-29 11:57 UTC (permalink / raw)
  To: buildroot

commit: https://git.buildroot.net/buildroot/commit/?id=86bc189c23c0ad4125460739b80c6b03d41a55ca
branch: https://git.buildroot.net/buildroot/commit/?id=refs/heads/master

The patches are removed as they are part of this release. 0.0.7 is a
bugfix and compatibility release to keep this usable for on newer systems.

Signed-off-by: Martin Kepplinger <martink@posteo.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
---
 .../0001-fix-segfault.patch                        |  34 ----
 .../0002-port-abi-12.patch                         | 196 ---------------------
 .../xdriver_xf86-input-tslib.hash                  |   4 +-
 .../xdriver_xf86-input-tslib.mk                    |   2 +-
 4 files changed, 3 insertions(+), 233 deletions(-)

diff --git a/package/x11r7/xdriver_xf86-input-tslib/0001-fix-segfault.patch b/package/x11r7/xdriver_xf86-input-tslib/0001-fix-segfault.patch
deleted file mode 100644
index adef18b..0000000
--- a/package/x11r7/xdriver_xf86-input-tslib/0001-fix-segfault.patch
+++ /dev/null
@@ -1,34 +0,0 @@
-Fetch from: https://launchpad.net/ubuntu/+archive/primary/+files/xf86-input-tslib_0.0.6-7build3.debian.tar.gz
-
-Signed-off-by: Jesper Baekdahl <jbb@gamblify.com>
----
---- xf86-input-tslib-0.0.6/src/tslib.c	2009-10-19 18:07:18.000000000 +0300
-+++ xf86-input-tslib-0.0.6.new/src/tslib.c	2010-02-12 16:15:05.000000000 +0200
-@@ -103,8 +103,6 @@
- static void
- PointerControlProc(DeviceIntPtr dev, PtrCtrl * ctrl)
- {
--	ErrorF("%s\n", __FUNCTION__);
--	return;
- }
- 
- static Bool
-@@ -406,7 +404,9 @@
- 		xf86MotionHistoryAllocate(pInfo);
- #endif
- 
--		break;
-+		if (!InitPtrFeedbackClassDeviceStruct(device, PointerControlProc))
-+			return !Success;
-+		break;
- 
- 	case DEVICE_ON:
- 		AddEnabledDevice(pInfo->fd);
-@@ -435,6 +435,7 @@
- 	xf86TslibControlProc(pInfo->dev, DEVICE_OFF);
- 	ts_close(priv->ts);
- 	xfree(pInfo->private);
-+	pInfo->private = NULL;
- 	xf86DeleteInput(pInfo, 0);
- }
- 
diff --git a/package/x11r7/xdriver_xf86-input-tslib/0002-port-abi-12.patch b/package/x11r7/xdriver_xf86-input-tslib/0002-port-abi-12.patch
deleted file mode 100644
index eb8fb54..0000000
--- a/package/x11r7/xdriver_xf86-input-tslib/0002-port-abi-12.patch
+++ /dev/null
@@ -1,196 +0,0 @@
-Fetch from: https://launchpad.net/ubuntu/+archive/primary/+files/xf86-input-tslib_0.0.6-7build3.debian.tar.gz
-Fixes build against newer versions of xorg.
-
-Signed-off-by: Jesper Baekdahl <jbb@gamblify.com>
----
-Index: xf86-input-tslib-trunk/src/tslib.c
-===================================================================
---- xf86-input-tslib-trunk/src/tslib.c	(revision 48)
-+++ xf86-input-tslib-trunk/src/tslib.c	(working copy)
-@@ -69,6 +69,13 @@
- #define DEFAULT_HEIGHT		240
- #define DEFAULT_WIDTH		320
- 
-+#if GET_ABI_MAJOR(ABI_XINPUT_VERSION) < 12
-+#define COLLECT_INPUT_OPTIONS(pInfo, options) xf86CollectInputOptions((pInfo), (options), NULL)
-+#else
-+#define COLLECT_INPUT_OPTIONS(pInfo, options) xf86CollectInputOptions((pInfo), (options))
-+#endif
-+
-+
- enum { TSLIB_ROTATE_NONE=0, TSLIB_ROTATE_CW=270, TSLIB_ROTATE_UD=180, TSLIB_ROTATE_CCW=90 };
- 
- enum button_state { BUTTON_NOT_PRESSED = 0, BUTTON_1_PRESSED = 1, BUTTON_3_CLICK = 3, BUTTON_3_CLICKED=4, BUTTON_EMULATION_OFF=-1 };
-@@ -106,7 +113,7 @@
- }
- 
- static Bool
--ConvertProc( LocalDevicePtr local,
-+ConvertProc( InputInfoPtr local,
- 			 int first,
- 			 int num,
- 			 int v0,
-@@ -135,7 +142,7 @@
-     return t;
- }
- 
--static void ReadInput (LocalDevicePtr local)
-+static void ReadInput (InputInfoPtr local)
- {
- 	struct ts_priv *priv = (struct ts_priv *) (local->private);
- 	struct ts_sample samp;
-@@ -382,7 +389,11 @@
- 					       axiswidth - 1,	/* max val */
- 					       axiswidth,	/* resolution */
- 					       0,		/* min_res */
--					       axiswidth);	/* max_res */
-+					       axiswidth	/* max_res */
-+#if GET_ABI_MAJOR(ABI_XINPUT_VERSION) >= 12
-+					       ,Absolute
-+#endif
-+					       );
- 
- 		InitValuatorAxisStruct(device, 1,
- #if GET_ABI_MAJOR(ABI_XINPUT_VERSION) >= 7
-@@ -392,7 +403,11 @@
- 					       axisheight - 1,	/* max val */
- 					       axisheight,	/* resolution */
- 					       0,		/* min_res */
--					       axisheight);	/* max_res */
-+					       axisheight	/* max_res */
-+#if GET_ABI_MAJOR(ABI_XINPUT_VERSION) >= 12
-+					       ,Absolute
-+#endif
-+					       );
- 
- 		if (InitProximityClassDeviceStruct (device) == FALSE) {
- 			ErrorF ("Unable to allocate EVTouch touchscreen ProximityClassDeviceStruct\n");
-@@ -434,7 +449,7 @@
- 	ErrorF("%s\n", __FUNCTION__);
- 	xf86TslibControlProc(pInfo->dev, DEVICE_OFF);
- 	ts_close(priv->ts);
--	xfree(pInfo->private);
-+	free(pInfo->private);
- 	pInfo->private = NULL;
- 	xf86DeleteInput(pInfo, 0);
- }
-@@ -444,47 +459,57 @@
-  *
-  * called when the module subsection is found in XF86Config
-  */
-+#if GET_ABI_MAJOR(ABI_XINPUT_VERSION) >= 12
-+static int 
-+xf86TslibInit(InputDriverPtr drv, InputInfoPtr pInfo, int flags)
-+#else
- static InputInfoPtr
- xf86TslibInit(InputDriverPtr drv, IDevPtr dev, int flags)
-+#endif
- {
- 	struct ts_priv *priv;
- 	char *s;
-+#if GET_ABI_MAJOR(ABI_XINPUT_VERSION) < 12
- 	InputInfoPtr pInfo;
-+#endif
- 
--	priv = xcalloc (1, sizeof (struct ts_priv));
-+	priv = calloc (1, sizeof (struct ts_priv));
-         if (!priv)
--                return NULL;
-+                return BadValue;
- 
-+#if GET_ABI_MAJOR(ABI_XINPUT_VERSION) < 12
- 	if (!(pInfo = xf86AllocateInput(drv, 0))) {
--		xfree(priv);
--		return NULL;
-+		free(priv);
-+		return BadValue;
- 	}
- 
- 	/* Initialise the InputInfoRec. */
- 	pInfo->name = dev->identifier;
--	pInfo->type_name = XI_TOUCHSCREEN;
- 	pInfo->flags =
- 	    XI86_KEYBOARD_CAPABLE | XI86_POINTER_CAPABLE |
- 	    XI86_SEND_DRAG_EVENTS;
--	pInfo->device_control = xf86TslibControlProc;
--	pInfo->read_input = ReadInput;
- #if GET_ABI_MAJOR(ABI_XINPUT_VERSION) == 0
- 	pInfo->motion_history_proc = xf86GetMotionEvents;
- 	pInfo->history_size = 0;
- #endif
--	pInfo->control_proc = NULL;
-+	pInfo->conf_idev = dev;
- 	pInfo->close_proc = NULL;
--	pInfo->switch_mode = NULL;
- 	pInfo->conversion_proc = ConvertProc;
- 	pInfo->reverse_conversion_proc = NULL;
--	pInfo->dev = NULL;
- 	pInfo->private_flags = 0;
- 	pInfo->always_core_feedback = 0;
--	pInfo->conf_idev = dev;
-+#endif
-+
-+	pInfo->type_name = XI_TOUCHSCREEN;
-+	pInfo->control_proc = NULL;
-+	pInfo->read_input = ReadInput;
-+	pInfo->device_control = xf86TslibControlProc;
-+	pInfo->switch_mode = NULL;
- 	pInfo->private = priv;
-+	pInfo->dev = NULL;
- 
- 	/* Collect the options, and process the common options. */
--	xf86CollectInputOptions(pInfo, NULL, NULL);
-+	COLLECT_INPUT_OPTIONS(pInfo, NULL);
- 	xf86ProcessCommonOptions(pInfo, pInfo->options);
- 
- 	priv->screen_num = xf86SetIntOption(pInfo->options, "ScreenNumber", 0 );
-@@ -510,23 +535,31 @@
- 		priv->rotate = TSLIB_ROTATE_NONE;
- 	}
- 
-+#if GET_ABI_MAJOR(ABI_XINPUT_VERSION) < 12
-  	s = xf86CheckStrOption(dev->commonOptions, "path", NULL);
-+#else
-+	s = xf86CheckStrOption(pInfo->options, "path", NULL);
-+#endif
-   	if (!s)
-+#if GET_ABI_MAJOR(ABI_XINPUT_VERSION) < 12
- 		s = xf86CheckStrOption(dev->commonOptions, "Device", NULL);
-+#else
-+		s = xf86CheckStrOption(pInfo->options, "Device", NULL);
-+#endif
-  
- 	priv->ts = ts_open(s, 1);
--	xfree(s);
-+	free(s);
- 
- 	if (!priv->ts) {
- 		ErrorF("ts_open failed (device=%s)\n",s);
- 		xf86DeleteInput(pInfo, 0);
--		return NULL;
-+		return BadValue;
- 	}
- 
- 	if (ts_config(priv->ts)) {
- 		ErrorF("ts_config failed\n");
- 		xf86DeleteInput(pInfo, 0);
--		return NULL;
-+		return BadValue;
- 	}
- 
- 	pInfo->fd = ts_fd(priv->ts);
-@@ -536,11 +569,13 @@
- 		priv->state = BUTTON_EMULATION_OFF;
- 	}
- 
-+#if GET_ABI_MAJOR(ABI_XINPUT_VERSION) < 12
- 	/* Mark the device configured */
- 	pInfo->flags |= XI86_CONFIGURED;
-+#endif
- 
- 	/* Return the configured device */
--	return (pInfo);
-+	return Success;
- }
- 
- _X_EXPORT InputDriverRec TSLIB = {
diff --git a/package/x11r7/xdriver_xf86-input-tslib/xdriver_xf86-input-tslib.hash b/package/x11r7/xdriver_xf86-input-tslib/xdriver_xf86-input-tslib.hash
index 9eb3f1d..b8eb2b8 100644
--- a/package/x11r7/xdriver_xf86-input-tslib/xdriver_xf86-input-tslib.hash
+++ b/package/x11r7/xdriver_xf86-input-tslib/xdriver_xf86-input-tslib.hash
@@ -1,2 +1,2 @@
-# Locally calculated
-sha256 5f46fdef095a6e44a69e0f0b57c7d665224b26d990d006611236d8332e85b105  xf86-input-tslib-0.0.6.tar.bz2
+# https://github.com/merge/xf86-input-tslib/releases/download/0.0.7/xf86-input-tslib-0.0.7.tar.bz2.sha256
+sha256 6f23cc9702b0ae16086d364b275335c094efbf6acde57f8a030e4db5b9aece03  xf86-input-tslib-0.0.7.tar.bz2
diff --git a/package/x11r7/xdriver_xf86-input-tslib/xdriver_xf86-input-tslib.mk b/package/x11r7/xdriver_xf86-input-tslib/xdriver_xf86-input-tslib.mk
index f8281d7..2e78634 100644
--- a/package/x11r7/xdriver_xf86-input-tslib/xdriver_xf86-input-tslib.mk
+++ b/package/x11r7/xdriver_xf86-input-tslib/xdriver_xf86-input-tslib.mk
@@ -4,7 +4,7 @@
 #
 ################################################################################
 
-XDRIVER_XF86_INPUT_TSLIB_VERSION = 0.0.6
+XDRIVER_XF86_INPUT_TSLIB_VERSION = 0.0.7
 XDRIVER_XF86_INPUT_TSLIB_SOURCE = xf86-input-tslib-$(XDRIVER_XF86_INPUT_TSLIB_VERSION).tar.bz2
 XDRIVER_XF86_INPUT_TSLIB_SITE = https://github.com/merge/xf86-input-tslib/releases/download/$(XDRIVER_XF86_INPUT_TSLIB_VERSION)
 XDRIVER_XF86_INPUT_TSLIB_LICENSE = MIT

^ permalink raw reply related

* [Buildroot] [git commit] x11r7: xdriver_xf86-input-tslib: new upstream location
From: Thomas Petazzoni @ 2017-04-29 11:56 UTC (permalink / raw)
  To: buildroot

commit: https://git.buildroot.net/buildroot/commit/?id=2ae6acd3c11c991cefa208d5005ee613c8408721
branch: https://git.buildroot.net/buildroot/commit/?id=refs/heads/master

This switches upstream to the Github project where xf86-input-tslib is
currently maintained - in cooperation with Pengutronix, who had hosted
the tarball release up until now.

Signed-off-by: Martin Kepplinger <martink@posteo.de>
[Thomas: fix XDRIVER_XF86_INPUT_TSLIB_SITE value.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
---
 package/x11r7/xdriver_xf86-input-tslib/Config.in                   | 2 +-
 package/x11r7/xdriver_xf86-input-tslib/xdriver_xf86-input-tslib.mk | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/package/x11r7/xdriver_xf86-input-tslib/Config.in b/package/x11r7/xdriver_xf86-input-tslib/Config.in
index 9c6e296..08d790c 100644
--- a/package/x11r7/xdriver_xf86-input-tslib/Config.in
+++ b/package/x11r7/xdriver_xf86-input-tslib/Config.in
@@ -8,7 +8,7 @@ config BR2_PACKAGE_XDRIVER_XF86_INPUT_TSLIB
 	help
 	  Touch screen library input driver
 
-	  http://www.ptxdist.org/software/xf86-input-tslib/index_en.html
+	  https://github.com/merge/xf86-input-tslib
 
 comment "xf86-input-tslib needs a toolchain w/ dynamic library"
 	depends on BR2_STATIC_LIBS
diff --git a/package/x11r7/xdriver_xf86-input-tslib/xdriver_xf86-input-tslib.mk b/package/x11r7/xdriver_xf86-input-tslib/xdriver_xf86-input-tslib.mk
index cd8c10a..f8281d7 100644
--- a/package/x11r7/xdriver_xf86-input-tslib/xdriver_xf86-input-tslib.mk
+++ b/package/x11r7/xdriver_xf86-input-tslib/xdriver_xf86-input-tslib.mk
@@ -6,7 +6,7 @@
 
 XDRIVER_XF86_INPUT_TSLIB_VERSION = 0.0.6
 XDRIVER_XF86_INPUT_TSLIB_SOURCE = xf86-input-tslib-$(XDRIVER_XF86_INPUT_TSLIB_VERSION).tar.bz2
-XDRIVER_XF86_INPUT_TSLIB_SITE = http://www.pengutronix.de/software/xf86-input-tslib/download
+XDRIVER_XF86_INPUT_TSLIB_SITE = https://github.com/merge/xf86-input-tslib/releases/download/$(XDRIVER_XF86_INPUT_TSLIB_VERSION)
 XDRIVER_XF86_INPUT_TSLIB_LICENSE = MIT
 XDRIVER_XF86_INPUT_TSLIB_LICENSE_FILES = COPYING
 XDRIVER_XF86_INPUT_TSLIB_DEPENDENCIES = \

^ permalink raw reply related

* [Buildroot] [PATCH v8 1/1] package/kodi: add optional support for pulseaudio
From: Thomas Petazzoni @ 2017-04-29 11:51 UTC (permalink / raw)
  To: buildroot
In-Reply-To: <20170429111410.29929-1-bernd.kuhls@t-online.de>

Hello,

On Sat, 29 Apr 2017 13:14:10 +0200, Bernd Kuhls wrote:
> Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
> ---
> v8: exteneded comment about libglib2 dependency and its role in Kodi
>     (Thomas)

Applied to master, thanks.

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

^ permalink raw reply

* [Buildroot] [git commit] package/kodi: add optional support for pulseaudio
From: Thomas Petazzoni @ 2017-04-29 11:47 UTC (permalink / raw)
  To: buildroot

commit: https://git.buildroot.net/buildroot/commit/?id=ff45194b3c61a5b88b3cac047b16ba85bef8a15e
branch: https://git.buildroot.net/buildroot/commit/?id=refs/heads/master

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
---
 package/kodi/Config.in | 11 +++++++++++
 package/kodi/kodi.mk   |  7 +++++++
 2 files changed, 18 insertions(+)

diff --git a/package/kodi/Config.in b/package/kodi/Config.in
index bf4ea01..fc40343 100644
--- a/package/kodi/Config.in
+++ b/package/kodi/Config.in
@@ -192,6 +192,17 @@ config BR2_PACKAGE_KODI_NONFREE
 	  Enable components with non-free licenses. This option
 	  currently only controls support for RAR archives.
 
+config BR2_PACKAGE_KODI_PULSEAUDIO
+	bool "pulseaudio"
+	depends on BR2_TOOLCHAIN_HAS_SYNC_4 # pulseaudio -> json-c
+	# Pulseaudio support in kodi needs glib support in Pulseaudio,
+	# see FindPulseAudio.cmake. Kodi meets all dependencies of
+	# libglib2, so there is no need to propagate them here.
+	select BR2_PACKAGE_LIBGLIB2
+	select BR2_PACKAGE_PULSEAUDIO
+	help
+	  Enable Pulseaudio support.
+
 config BR2_PACKAGE_KODI_LIBSHAIRPLAY
 	bool "shairport"
 	select BR2_PACKAGE_LIBSHAIRPLAY
diff --git a/package/kodi/kodi.mk b/package/kodi/kodi.mk
index a4835f0..12dc311 100644
--- a/package/kodi/kodi.mk
+++ b/package/kodi/kodi.mk
@@ -333,6 +333,13 @@ else
 KODI_CONF_OPTS += -DENABLE_OPTICAL=OFF
 endif
 
+ifeq ($(BR2_PACKAGE_KODI_PULSEAUDIO),y)
+KODI_CONF_OPTS += -DENABLE_PULSEAUDIO=ON
+KODI_DEPENDENCIES += pulseaudio
+else
+KODI_CONF_OPTS += -DENABLE_PULSEAUDIO=OFF
+endif
+
 # Remove versioncheck addon, updating Kodi is done by building a new
 # buildroot image.
 KODI_ADDON_MANIFEST = $(TARGET_DIR)/usr/share/kodi/system/addon-manifest.xml

^ permalink raw reply related

* [Buildroot] [PATCH 1/1] package/kodi-visualisation-goom: bump version
From: Thomas Petazzoni @ 2017-04-29 11:46 UTC (permalink / raw)
  To: buildroot
In-Reply-To: <20170429105544.23417-1-bernd.kuhls@t-online.de>

Hello,

On Sat, 29 Apr 2017 12:55:44 +0200, Bernd Kuhls wrote:
> This bump fixes a compile error on powerpc.
> 
> Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
> ---
> I forgot to add this patch to v7 of the patch series, sorry.
> My fix was committed upstream today.

Applied to master, thanks.

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

^ permalink raw reply

* [Buildroot] [git commit] package/kodi-visualisation-goom: bump version
From: Thomas Petazzoni @ 2017-04-29 11:46 UTC (permalink / raw)
  To: buildroot

commit: https://git.buildroot.net/buildroot/commit/?id=a29a0c96194aaf12b482f76486c19479d5769c2d
branch: https://git.buildroot.net/buildroot/commit/?id=refs/heads/master

This bump fixes a compile error on powerpc.

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
---
 package/kodi-visualisation-goom/kodi-visualisation-goom.hash | 2 +-
 package/kodi-visualisation-goom/kodi-visualisation-goom.mk   | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/package/kodi-visualisation-goom/kodi-visualisation-goom.hash b/package/kodi-visualisation-goom/kodi-visualisation-goom.hash
index 8d4548a..fa5fbf6 100644
--- a/package/kodi-visualisation-goom/kodi-visualisation-goom.hash
+++ b/package/kodi-visualisation-goom/kodi-visualisation-goom.hash
@@ -1,2 +1,2 @@
 # Locally computed
-sha256 62a26efb6cd5d5fd109e38accb3deb71145972ae30b914bed7620eb40465c954  kodi-visualisation-goom-v1.1.0.tar.gz
+sha256 501b8b1d5bd174ec71d57894293ae5eb0c1f41c3b725ab5e7b5a1c5eeea77a7f  kodi-visualisation-goom-54da35870930acd1a3a36195dd2c1498ac336b90.tar.gz
diff --git a/package/kodi-visualisation-goom/kodi-visualisation-goom.mk b/package/kodi-visualisation-goom/kodi-visualisation-goom.mk
index 48e0425..a7503a9 100644
--- a/package/kodi-visualisation-goom/kodi-visualisation-goom.mk
+++ b/package/kodi-visualisation-goom/kodi-visualisation-goom.mk
@@ -4,7 +4,7 @@
 #
 ################################################################################
 
-KODI_VISUALISATION_GOOM_VERSION = v1.1.0
+KODI_VISUALISATION_GOOM_VERSION = 54da35870930acd1a3a36195dd2c1498ac336b90
 KODI_VISUALISATION_GOOM_SITE = $(call github,notspiff,visualization.goom,$(KODI_VISUALISATION_GOOM_VERSION))
 KODI_VISUALISATION_GOOM_LICENSE = GPL-2.0+
 KODI_VISUALISATION_GOOM_LICENSE_FILES = src/Main.cpp

^ permalink raw reply related


This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox