Age | Commit message (Collapse) | Author |
|
Since this g++.conf file was added to fix some MX6 issues the upstream
version of the file in which it's based has received some more fixes:
http://git.openembedded.org/openembedded-core/commit/?id=6456b596a73232429e0d223089f32dc350ea20b0
http://git.openembedded.org/openembedded-core/commit/?id=6c78f629323bea0be44181d60e3ce45be0416093
This commit fixes an error building QT applications in debug mode
(CONFIG+=debug in the QT project file) with a toolchain generated for a
MX6 machine (with populate_sdk).
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
libdrm build fails due to libdrm_%.bbappend, since 'SRC_URI_append'
doesn't contain leading space in appended string.
Signed-off-by: Alexander Shashkevich <alex@stunpix.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Some patches were not removed while updating the recipe. This dropes
the left over files to remove them from the layer.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
The U-Boot mainline uses 69K as offset so we ought to use it here;
this aligns the SD card image layout to work out of box with all
U-Boot mainline based SPL boards.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Using 'cp -a' to install files in do_install function leaks user/group
attributes from build machine to target packages.
WARNING: log_check: There is a warn message in the logfile
WARNING: log_check: Matched keyword: [warn]
WARNING: log_check: warning: user xxxxx does not exist - using root
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
as long as the patch applies chances are high that we need it
Signed-off-by: Andreas Müller <schnitzeltony@googlemail.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
In case we have kernel modules, a runtime dependency or recommendation
can break the dependency resolution algorithm of BitBake resulting on
a conflict like:
,----[ Provider conflict error ]
| ERROR: Trying to resolve runtime dependency kernel-module-nfs resulted
| in conflicting PREFERRED_PROVIDER entries being found.
| The providers found were:
| ['.../meta-customer/recipes-kernel/linux/linux-customer-mfgtool_3.10.17.bb',
| '.../meta-customer/recipes-kernel/linux/linux-customer_3.10.17.bb']
| The PREFERRED_PROVIDER entries resulting in this conflict were:
| ['PREFERRED_PROVIDER_linux-mfgtool = mfgtool-linux-customer',
| 'PREFERRED_PROVIDER_virtual/kernel = linux-customer']
`----
To fix this, we need to rename the dynamic packages and hence change
the built binaries, solving the conflict.
Change-Id: I8becb8a4a839494720711f60f51515aba0140c68
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
This updates the patches-3.14 branch to 3.14.19 stable release.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
This includes following fixes:
9e76b38 wandboard: Use 32bit color depth for Fusion LCD
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Changes:
- ipu, vpu, v4l2src: Use reference counting for resources
- ipu: use a global FD instead of per-blitter FDs
- introduce a proper physical address type and a printf conversion
specifier
- remove cpu address from GstImxPhysMemory and add an "internal" pointer
- replace obsolete "padding" member in phys mem meta with better defined
- use a reference counter for mapping physical memory
- ipu: fix kernel version detection and include paths in wscript
- eglvivsink: remove RGB from the list of supported formats
- cleanup & logging improvements
Signed-off-by: Carlos Rafael Giani <dv@pseudoterminal.org>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
This quiets GNU_HASH warnings when building with an external toolchain.
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Changes:
- Fixed imxipuvideotransform's passthrough mode to make it work with
playbin's about-to-finish signal
- Display scale updates in imxeglvivsink weren't being handled properly
- Updates to make imxv4l2src buffers shareable
- Extra calls to make imxv4l2src work for TV capture
- Better zero-copy support in tee'd pipelines
- New IDR interval property in h264 encoder
- New slice size, and intra refresh properties in encoders
- Configurable framebuffer number in IPU sink
- Workaround for GStreamer 1.4.0 bug related to the
GST_BUFFER_FLAG_TAG_MEMORY buffer flag
- Unset GST_BUFFER_FLAG_TAG_MEMORY explicitely in VPU decoder to prevent
unnecessary copies
- Drain decoder when a new format is set, allowing for seamless transitions
- Fixed decoder shutdown freeze
- Fixed leaked socket pair problem in imxeglvivsink
Signed-off-by: Carlos Rafael Giani <dv@pseudoterminal.org>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Freescale when doing 3.10.17-1.0.1 has ignored the fixes done, by
themselves, in the 3.10.17-1.0.0 kernel.
This includes following patches, rebased on top 3.10.17-1.0.1:
ENGR00318392 ARM:imx6x: Save/restore SCU and some CP15 registers across suspend/resume
ENGR00301078-2: ARM: dts: imx6sl-evk: add support for pfuze200 on imx6sl-evk
ENGR00301078-1: ARM: dts: imx6dl-sabresd: add support for pfuze200 on mx6dl-sabresd
pfuze100-regulator: Fix of_node_get() parameter
regulator: pfuze100: Use of_get_child_by_name
regulator: pfuze100: Add PFUZE200 support to Kconfig and module description
regulator: pfuze100: Add terminate entry for [i2c|of]_device_id tables
regulator: pfuze100: add pfuze200 support
regulator: pfuze100: Fix address of FABID
ENGR00309838 ARM: imx6sl: gpc: add chip revision check for dispmix
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Since CPU frequency scaling is a bad idea for real-time systems anyway,
use 'performance' and disable the remaining cpufreq governors.
Avoids the following issue:
BUG: scheduling while atomic: swapper/0/0/0x00000002
Modules linked in:
Preemption disabled at:[< (null)>] (null)
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 3.10.17-rt12-monkey+g232293e #1
[<80013c7c>] (unwind_backtrace+0x0/0xec) from [<80011644>] (show_stack+0x10/0x14)
[<80011644>] (show_stack+0x10/0x14) from [<8004e160>] (__schedule_bug+0x60/0xa8)
[<8004e160>] (__schedule_bug+0x60/0xa8) from [<80643458>] (__schedule+0x408/0x4a8)
[<80643458>] (__schedule+0x408/0x4a8) from [<8064352c>] (schedule+0x34/0xa4)
[<8064352c>] (schedule+0x34/0xa4) from [<80644104>] (rt_spin_lock_slowlock+0xc0/0x298)
[<80644104>] (rt_spin_lock_slowlock+0xc0/0x298) from [<80033780>] (lock_timer_base.isra.29+0x2c/0x4c)
[<80033780>] (lock_timer_base.isra.29+0x2c/0x4c) from [<80033944>] (mod_timer+0x60/0x1c4)
[<80033944>] (mod_timer+0x60/0x1c4) from [<8044bcc8>] (cpufreq_interactive_idle_notifier+0x148/0x180)
[<8044bcc8>] (cpufreq_interactive_idle_notifier+0x148/0x180) from [<8004ad40>] (notifier_call_chain+0x44/0x84)
[<8004ad40>] (notifier_call_chain+0x44/0x84) from [<8004afd4>] (atomic_notifier_call_chain+0x34/0x48)
[<8004afd4>] (atomic_notifier_call_chain+0x34/0x48) from [<8005d9a0>] (cpu_startup_entry+0xfc/0x164)
[<8005d9a0>] (cpu_startup_entry+0xfc/0x164) from [<80c1cab4>] (start_kernel+0x324/0x330)
[<80c1cab4>] (start_kernel+0x324/0x330) from [<10008074>] (0x10008074)
See:
https://lkml.org/lkml/2014/6/10/457
https://lists.yoctoproject.org/pipermail/meta-freescale/2014-June/008957.html
Signed-off-by: Jacob Kroon <jacob.kroon@mikrodidakt.se>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Signed-off-by: Jacob Kroon <jacob.kroon@mikrodidakt.se>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
This reverts commit 4ffc6ac2dc2f3390a95d125403b249f674d13c9d.
This change breaks the SDK build. Until the nativesdk/native fix
is backported to Daisy we cannot include this change.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
This includes the following main changes:
* Increase max code size of MQX binary that can be loaded to SRAM from
128KB to 256KB.
* Add compatibility for the 3.13+ Linux kernel. Functionality with the 3.0
kernel remains unchanged.
* Integrate patch that removes bashisms from kernel-module-mcc's Makefile.
Signed-off-by: Anthony Felice <tony.felice@timesys.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
This version was released to integrate the patch that removed bashisms from
the libmcc Makefile.
Signed-off-by: Anthony Felice <tony.felice@timesys.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
For 2D GPU cores there is a potential need to manage the maximum number
of outstanding return requests from the memory subsystem. The proper way
to manage these for 2D cores with this configuration is to use a FLUSH.
The patch does a FLUSH after every DRAW, the pipe will not overfill and
the scenario which may lead to a stall will not occur.
Signed-off-by: Lauren Post <lauren.post@freescale.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
For 2D GPU cores there is a potential need to manage the maximum number
of outstanding return requests from the memory subsystem. The proper way
to manage these for 2D cores with this configuration is to use a FLUSH.
The patch does a FLUSH after every DRAW, the pipe will not overfill and
the scenario which may lead to a stall will not occur.
Signed-off-by: Lauren Post <lauren.post@freescale.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
For 2D GPU cores there is a potential need to manage the maximum number
of outstanding return requests from the memory subsystem. The proper way
to manage these for 2D cores with this configuration is to use a FLUSH.
The patch does a FLUSH after every DRAW, the pipe will not overfill and
the scenario which may lead to a stall will not occur.
This patch requires the Linux kernel 3.10.17-1.0.1 fixes.
Signed-off-by: Lauren Post <lauren.post@freescale.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Signed-off-by: Mario Domenech Goulart <mario@ossystems.com.br>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Replacing bash brace expansion with a Posix compliant syntax to make
do_install succeed also when using a dash shell.
Signed-off-by: Max Krummenacher <max.oss.09@gmail.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Replacing bash brace expansion with a Posix compliant syntax to make
do_install succeed also when using a dash shell.
Signed-off-by: Max Krummenacher <max.oss.09@gmail.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
The Vivante DRI library requires the libdrm to proper work; this adds
the missing dependency.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Pulseaudio crashes when other program using sound:
Assertion 'pthread_mutex_unlock(&m->mutex) == 0' failed
at pulsecore/mutex-posix.c:106, function pa_mutex_unlock().
Disable the feature PTHREAD priority inheritance is the easiest
workaround for this issue.
Signed-off-by: Jun Zhu <R01007@freescale.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Signed-off-by: Wang Zidan <b50113@freescale.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
This package couldn't build using a toolchain with kernel headers >=
3.10 because of the use of VIDIOC_DBG_G_CHIP_IDENT. As stated in
videodev2.h, this ioctl should never be used.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
This fixes the following build error:
,----
| .../configure: line 4229: syntax error near unexpected token `DIRECTFB,'
| .../configure: line 4229: `PKG_CHECK_MODULES(DIRECTFB, directfb >= $DIRECTFB_REQUIRED_VERSION, WATER=`$PKG_CONFIG --variable=water directfb` ,'
`----
Change-Id: Ibd5807b1497ff83b4241cd53f97d18b62a0cb2df
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
This updates the Linux 3.14-based branch to the 3.14.17 stable
release.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
This updates the Linux 3.14-based branch to the 3.14.16 stable
release.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
This updates the Linux 3.14-based branch to the 3.14.14 stable
release.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
This includes following updates:
b3912bb ENGR00285448 [MX28] SAIF: Out of Voltage in SAIF clock PINs
90ce77d ENGR00285446-3 [MX28] SAIF: Bit Shift in SAIF RX Data
1ea685a ENGR00285446-2 [MX28] SAIF: Bit Shift in SAIF RX Data
1ca8992 ENGR00285446-1 [MX28] SAIF: Bit Shift in SAIF RX Data
45fe172 mach-mx5/devices.c: Add PMU device
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
The libgl-mx6 package had a duplicated libGL.so.1.
Force it into a symlink.
This was tested by comparing the output of "ls -lR --time-style=+ image/"
before and after the change.
Signed-off-by: Jacob Kroon <jacob.kroon@mikrodidakt.se>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
The do_install() function is run under pseudo in order to create packages with
correct file-permissions. By issuing "cp -a" to install files, we leak
build user/group permissions into the packages, since the do_unpack() function
is executed as the regular build user. Replace it with -P, so that we at least
keep symlinks.
Also drop -x (--one-file-system) since it's not needed.
This was tested by comparing the output of "ls -lR --time-style=+ image/"
before and after the change.
Signed-off-by: Jacob Kroon <jacob.kroon@mikrodidakt.se>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Running
readelf --dynamic libgc_wayland_protocol.so | grep NEEDED
shows that the library has a runtime dependency on libffi. Same thing
applies to libwayland-viv.so.
For non-wayland builds, where libffi may or may not be available during
package time, skip packaging them so that they don't cause unexpected
dependency changes in buildhistory.
Signed-off-by: Jacob Kroon <jacob.kroon@mikrodidakt.se>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
This updates the 3.14-based kernel to the 3.14.11 release.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
This includes the following updates to the 3.10.17-1.0.0-GA:
2a69800 ENGR00318392 ARM:imx6x: Save/restore SCU and some CP15
2a7bd41 ENGR00301078-2: ARM: dts: imx6sl-evk: add support for pfuze200 on imx6sl-evk
f0790e4 ENGR00301078-1: ARM: dts: imx6dl-sabresd: add support for pfuze200 on mx6dl-sabresd
d52df5f pfuze100-regulator: Fix of_node_get() parameter
ce1917e regulator: pfuze100: Use of_get_child_by_name
097577d regulator: pfuze100: Add PFUZE200 support to Kconfig and module description
dad7af3 regulator: pfuze100: Add terminate entry for [i2c|of]_device_id tables
792bf11 regulator: pfuze100: add pfuze200 support
df8c0fa regulator: pfuze100: Fix address of FABID
dd27604 ENGR00309838 ARM: imx6sl: gpc: add chip revision check for dispmix
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Changes:
- Support newer Vivante drivers, which use a different name for the
direct textures extension
- Sync points for forced keyframes during encoding were missing
- Added checks for cases where physical memory allocators fail
Signed-off-by: Carlos Rafael Giani <dv@pseudoterminal.org>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
This defines NAND parameters for u-boot.mxsboot-nand image creation
similar to those used for UBI image (for K9LBG08U0D-PCB chip):
- page size: 4 k
- OOB size: 218 bytes
- erase size: 512 k
Signed-off-by: Alexey Ignatov <lexszero@gmail.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
This adds support to generate the U-Boot image for use with NAND
(uboot.mxsboot-nand) and allows for custom params to be given for
mxsboot using the MXSBOOT_NAND_ARGS variable which should be defined
in the machine definition file.
Signed-off-by: Alexey Ignatov <lexszero@gmail.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
This character causes touch issues with calibration failing
Signed-off-by: Lauren Post <lauren.post@freescale.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
IOGetVirtMem() returns a pointer or specifically -1 (MAP_FAILED)
in the case of failure, and a couple of calls were improperly testing
for zero.
Signed-off-by: Eric Nelson <eric.nelson@boundarydevices.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Without this, the enlightenment foundation libraries will fail to build,
because Vivante EGL headers require these extra preprocessor defines
Signed-off-by: Carlos Rafael Giani <dv@pseudoterminal.org>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Without these, the package will not be built with the correct EGL
types and definitions
Signed-off-by: Carlos Rafael Giani <dv@pseudoterminal.org>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Signed-off-by: Carlos Rafael Giani <dv@pseudoterminal.org>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
Building with Angstrom toolchain fails on inline assembler added by
drm-update-arm.patch because of missing ite instruction. Insert
instruction to make the inline assembler block Thumb2 compatible.
Signed-off-by: Thomas Elste <thomas.elste@imms.de>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|
|
The IOGetVirtMem() routine returns the address of a memory block
or -1 (MAP_FAILED) to indicate an error.
Many callers of this routine tested the return value for <= 0
to detect failure, and at least with a 3G/1G memory split
configured in the kernel, a negative number (!= -1) is a
valid (successful) return value.
Without this patch, the IOSystemInit() will often detect
failure incorrectly.
Change-Id: I8b20ae5f74608307cbb810168024e8863599dbdb
Signed-off-by: Eric Nelson <eric.nelson@boundarydevices.com>
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
|