Tom Hochstein [Tue, 13 Dec 2016 17:55:44 +0000 (11:55 -0600)]
weston: Fix build on i.MX 6SL
Weston requires egl and glesv2 together, or neither. Since the i.MX 6SL
can only provide egl, we configure weston to work without egl. A build
break occurred due to i.MX patches which should not be applied for 6SL.
Signed-off-by: Tom Hochstein <tom.hochstein@nxp.com> Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Sona Sarmadi [Wed, 30 Nov 2016 12:17:39 +0000 (13:17 +0100)]
linux-qoriq: fix CVE-2016-0758
Fixes a flaw in the Linux kernel's ASN.1 DER decoder processed
certain certificate files with tags of indefinite length. A local,
unprivileged user could use a specially crafted X.509 certificate
DER file to crash the system or, potentially, escalate their
privileges on the system.
References:
https://lkml.org/lkml/2016/5/12/270
Upstream patch:
https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/patch/
?id=af00ae6ef5a2c73f21ba215c476570b7772a14fb [backported from stable 3.16]
Signed-off-by: Sona Sarmadi <sona.sarmadi@enea.com> Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Sona Sarmadi [Wed, 30 Nov 2016 12:17:38 +0000 (13:17 +0100)]
linux-qoriq: fix CVE-2016-2053
The asn1_ber_decoder function in lib/asn1_decoder.c in the Linux
kernel before 4.3 allows attackers to cause a denial of service
(panic) via an ASN.1 BER file that lacks a public key, leading
to mishandling by the public_key_verify_signature function in
crypto/asymmetric_keys/public_key.c.
Dominic Sacré [Mon, 21 Nov 2016 15:53:25 +0000 (16:53 +0100)]
linux-fslc-imx-rt: Add patch to support external imx-gpu-viv driver
Export functions swait_prepare_locked() and swait_finish_locked() to
make them available to the external imx-gpu-viv driver module.
Remove the check whether MACHINE_USES_VIVANTE_KERNEL_DRIVER_MODULE is
set; both the built-in driver and the external module now work with the
RT kernel.
Signed-off-by: Dominic Sacré <dominic.sacre@gmx.de> Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Tom Hochstein [Sat, 19 Nov 2016 01:11:49 +0000 (19:11 -0600)]
weston: Use compiler settings provided by IMX-GPU driver pkg-config
The IMX-GPU driver now provides compiler settings through pkg-config,
so don't set them explicitly. The package had to be patched to
link the compositor to egl.
Signed-off-by: Tom Hochstein <tom.hochstein@nxp.com> Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Fabio Berton [Thu, 17 Nov 2016 17:22:24 +0000 (15:22 -0200)]
linux-fslc: Enable TRACER support
Fix warning:
WARNING: lttng-modules-2.8.0+gitAUTOINC+6e4fc6f36d-r0 do_package:
lttng-modules: no modules were created; this may be due to
CONFIG_TRACEPOINTS not being enabled in your kernel.
Signed-off-by: Fabio Berton <fabio.berton@ossystems.com.br> Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Fabio Berton [Thu, 17 Nov 2016 17:22:23 +0000 (15:22 -0200)]
linux-imx: Enable TRACER support
Fix warning:
WARNING: lttng-modules-2.8.0+gitAUTOINC+6e4fc6f36d-r0 do_package:
lttng-modules: no modules were created; this may be due to
CONFIG_TRACEPOINTS not being enabled in your kernel.
Signed-off-by: Fabio Berton <fabio.berton@ossystems.com.br> Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Andreas Müller [Tue, 15 Nov 2016 15:30:07 +0000 (16:30 +0100)]
qtbase: do not pass -DEGL_API_FB=1 for x11 in distro features
* sympthoms:
* qml applications complain/crash with 'Failed to create OpenGL context...'
* qt5-opengles2-test hangs forever without displaying anything
* analysis:
qtbase's configure complains:
<...>/usr/include/EGL/eglvivante.h:117:16: note: class type '_FBDisplay' is incomplete
and configure ends up with no X11 support:
EGLFS ................ no
EGLFS i.MX6 ........ no
EGLFS i.MX6 Wayland. no
EGLFS EGLDevice .... no
EGLFS GBM .......... no
EGLFS Mali ......... no
EGLFS Raspberry Pi . no
EGLFS X11 .......... no
LinuxFB .............. no
Mir client............ no
XCB .................. yes (system library)
EGL on X ........... no
qtbase: Allow to build eglfs with x11 in DISTRO_FEATURES
* even when x11 is in DISTRO_FEATURES someone might configure qtbase
to build with eglfs, that works OK with Qt 5.5.1, because
-DEGL_API_FB=1 is also passed through
mkspecs/devices/linux-imx6-g++/qmake.conf
but with 5.6 and 5.7 it fails when X11 headers are include with their
own Status enum (the same errors like before
https://codereview.qt-project.org/#/c/105116/2)
Adding -DEGL_API_FB=1 like linux-imx6-g++/qmake.conf does is simplest
workaround to prevent imx-gpu-viv from including Xlib from
EGL/eglvivante.h
* comments on original commit (at today's' environment):
* it breaks GL support for X11
* in meta-freescale's qtbase.bbappend eglfs is disabled explicitly when
x11 in DISTRO_FEATURES so it seems that vivante driver does not support
eglfs and x11 at the same build.
Tested: setting 'eglfs' in PACKAGECONFIG builds fine but bare eglfs
support is skipped.
Signed-off-by: Andreas Müller <schnitzeltony@googlemail.com> Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
- 615fdbc Merge remote-tracking branch 'origin/master' into 2016.11+fslc
- 38cacda Merge branch 'master' of git://git.denx.de/u-boot-tegra
- a8d0526 ARM: tegra186: call secure monitor for all cache-wide ops
- 1ab557a armv8: add hooks for all cache-wide operations
- b9ae641 ARM: tegra: translate __asm_flush_l3_cache to assembly
- 6db8e17 ARM: tegra: ensure nvtboot_boot_x0 alignment
- d8c4eb6 Merge branch 'master' of git://git.denx.de/u-boot-net
- aa555fe net: use random ethernet address if invalid and not zero
- 6ecf9e2 net: mvgbe: Fix build error with CONFIG_PHYLIB
- f018545 net: phy: micrel: center FLP burst timing at 16ms
- 2162245 ARM: tegra: enable Ethernet on p2771-0000
- 31c1ff9 ARM: tegra: add DWC EQoS (ethernet) to Tegra186 DT
- 2b950f3 ARM: tegra: configure Ethernet address on Tegra186
- 86919a2 ARM: tegra: add SoC-level hook for board_late_init()
- ba4dfef net: add driver for Synopsys Ethernet QoS device
- afb970f dt: net: add DWC EQoS binding
- dac09fc wandboard: Make Ethernet functional again
- 0f7c6cd mkimage: Allow including a ramdisk in FIT auto mode
- baade49 travis-ci: Try harder to build all ARM targets
- e0f2406 buildman: Fix building based on 'options' field
- 4b6035d mx6sabresd: Make Ethernet functional again
- 97bb1f0 engicam: icorem6: Fix config files
- 68fd5c1 armv8: define get_ticks() for the ARMv8 Generic Timer
- 8add679 doc: update README.arm64
- 2a380cc tools: imximage: check return value when open the plugin file
- ae3de0d image: Protect against overflow in unknown_msg()
- eb9e699 colibri_pxa270: drop lzma support for space reason
- eef55e5 Merge branch 'master' of git://git.denx.de/u-boot-usb
- dd93a8e dfu: align array in dfu_get_dev_type with enum dfu_device_type
- dbdc274 cmd: dfu: Add error handling for board_usb_init
- 7fd1173 Merge git://git.denx.de/u-boot-rockchip
- 4594ac0 rk3288: kconfig: remove duplicate definition of SPL_MMC_SUPPORT
- 3f3e1e3 rockchip: doc: add GPT partition layout
- 7f35bbb rockchip: use rockchip linux partitions layout
- 612a2a0 dts: rk3288: remove node in dmc which not need anymore
- 7d6c78f rk3288: sdram: auto-detect the capacity
- cd2f6b8 rk3288: config change for enable dram capacity auto-detect.
- 0717dde evb-rk3399: config: set emmc as default boot dev
- 1b64a05 rk3288: fix reg address for GRF_SOC_CON2
- 27b95d2 rk3399: disable the clock multiplier support when SoC init
- 5564ed5 rockchip: rk3288: Move rockchip_get_cru() out of the driver
- c8a6bc9 rockchip: rk3399: Move rockchip_get_cru() out of the driver
- 92ac73e rockchip: rk3036: Move rockchip_get_cru() out of the driver
- 8dfb4a2 clk: rk3399: fix rockchip_get_cru
- 6016982 rockchip: RK3288 needs fdt and initrd below 256M now
Signed-off-by: Fabio Berton <fabio.berton@ossystems.com.br> Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
- properly pass on color format in simplified JPEG encoder interface
- add alternative write-callback-style encoding mode
also add encode example variant which uses write-callback style output
- add support for "fake grayscale mode" in encoders
this is done by using I420 internally and filling the U and V planes
with 0x80 bytes
- make sure JPEG quantization table is copied in standardized zig zag order
the VPU does not, so this has to be done explicitely
Signed-off-by: Carlos Rafael Giani <dv@pseudoterminal.org> Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
imx-gst1.0-plugin: configure for available platforms only
For the MACHINEOVERRIDES "mx6dl" and "mx6q" PLATFORM is set to MX6DL and MX6Q,
respectively, resulting in the following error from the configure script:
| configure: error: No target platform specified!
| Use ./configure PLATFORM=(MX6/MX6QP/MX6SL/MX6SX/MX6UL/MX7D) to retry
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com> Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Richard Purdie [Wed, 2 Nov 2016 12:11:06 +0000 (12:11 +0000)]
linux-mgftool: Fix taskhash changed error
We need to expand KERNEL_VERSION at runtime, not at parse time. The
class extension code can expand variables early so remove the variable
from the datastore temporarily.
Change-Id: Ic0c8c8f62447e8368ce19c2420812f75009a295f Reported-by: Jun Zhu <junzhu@nxp.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org> Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Adrian Dudau [Mon, 31 Oct 2016 14:53:01 +0000 (15:53 +0100)]
linux-qoriq: fix CVE-2016-5696
net/ipv4/tcp_input.c in the Linux kernel before 4.7 does not
properly determine the rate of challenge ACK segments, which
makes it easier for man-in-the-middle attackers to hijack TCP
sessions via a blind in-window attack.
https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/
patch/?id=72c2d3bccaba4a0a4de354f9d2d24eccd05bfccf (This a follow-up
to "tcp: make challenge acks less predictable)
Signed-off-by: Adrian Dudau <adrian.dudau@enea.com> Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Fabio Berton [Thu, 27 Oct 2016 11:50:31 +0000 (09:50 -0200)]
linux-fslc-imx: Enable TRACER support
Fix warning:
WARNING: lttng-modules-2.8.0+gitAUTOINC+6e4fc6f36d-r0 do_package:
lttng-modules: no modules were created; this may be due to
CONFIG_TRACEPOINTS not being enabled in your kernel.
Signed-off-by: Fabio Berton <fabio.berton@ossystems.com.br> Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Yuqing Zhu [Thu, 13 Oct 2016 01:59:54 +0000 (09:59 +0800)]
gstreamer1.0-plugins-bad: Fix MRT cannot work in GLES3.0
1. Add #ifdefine Marco to avoid redefinition of GL_COLOR_ATTACHMENT(1,2).
2. Change to use glBlitFramebuffer for textures copy. glCopyTexImage2D
cannot work because we cannot get texture internal format.
Signed-off-by: Yuqing Zhu <carol.zhu@nxp.com> Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Yuqing Zhu [Thu, 13 Oct 2016 01:59:51 +0000 (09:59 +0800)]
gstreamer1.0-plugins-bad: Accelerate gldownload with directviv API
1) Propose a physical buffer pool to upstream in gldownload
2) Bind the physical buffer with texture via dirctviv
3) In gldownload, wrap the physical buffer to gstbuffer, pass to
downstream plugins.
Signed-off-by: Yuqing Zhu <carol.zhu@nxp.com> Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Yuqing Zhu [Thu, 13 Oct 2016 01:59:50 +0000 (09:59 +0800)]
gstreamer1.0-plugins-bad: Add directviv to glimagesink to improve playback performance
1.Add a physical buffer uploader in glupload plugin and using viv direct
texture to bind physical continious buffer with texture to avoid memory
copy from videobuffer to texture to gain good performance.
2.Reduce glimagesink load latency by override glimagesink ALLOCATION query to
avoid memory copy.
Signed-off-by: Yuqing Zhu <carol.zhu@nxp.com> Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
