Set the romley preferred kernel version to 3.4.

Signed-off-by: Kishore Bodke <kishore.k.bodke@intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com>

Add linux-yocto_3.4 kernel support for the romley machine.

Signed-off-by: Kishore Bodke <kishore.k.bodke@intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com>

Update the MAINTAINERS list after removing the cedartrail BSP from
meta-intel.

Signed-off-by: Kishore Bodke <kishore.k.bodke@intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com>

Cedartrail will not be supported in the Yocto 1.4 Release.

Remove it from meta-intel.

Signed-off-by: Kishore Bodke <kishore.k.bodke@intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com>

RDEPENDS should be set on specific packages, not applied to everything
in general (e.g. -dev, -dbg or -doc packages). This fixes up warnings in
the build.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com>

Adjust bbappend to match changes in OE-Core.

Signed-off-by: Darren Hart <dvhart@linux.intel.com>

Using AUTOREV turns out to cause problems with unreproducible
autobuilder builds and unwanted network access.  This patch removes
kernel AUTOREVs and locks down SRCREVs and LINUX_VERSIONs.

Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com>

The recently added 1.16 version of the EMGD graphics stack also works
for the FRI2 BSP.

Update the BSP to use the newer EMGD version for graphics.

This partly addresses:
[YOCTO #3809]

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Acked-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

Mention the use of the graphics stack based on the kernel 'gma500'
driver as well as the 'modesetting' X driver for the emenlow-noemgd
machine configuration.

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

The following was seen when booting the fri2:

BUG: Bad page state in process loop0  pfn:3803
page:f6991700 count:0 mapcount:0 mapping:  (null) index:0x0
page flags: 0x80400000(uncached)
Modules linked in: emgd drm_kms_helper ip_tables x_tables arc4 minix cdc_acm iwlwifi led_class mac80211 cfg80211 firmware_class
Pid: 448, comm: loop0 Not tainted 3.4.26-yocto-standard #1
Call Trace:
 [<c10cdb0f>] bad_page+0x8f/0xe0
 [<c10ce0e5>] get_page_from_freelist+0x435/0x4e0
 [<c10ce2c5>] __alloc_pages_nodemask+0x135/0x690
 [<c10d1964>] __do_page_cache_readahead+0xe4/0x210
 [<c1055a6b>] ? get_parent_ip+0xb/0x40
 [<c1055a6b>] ? get_parent_ip+0xb/0x40
 [<c10d1cf7>] ra_submit+0x27/0x30
 [<c10d1e37>] ondemand_readahead+0x137/0x230
 [<c166b061>] ? __wait_on_bit_lock+0x61/0xa0
 [<c10d1fae>] page_cache_async_readahead+0x7e/0xa0
 [<c112b590>] __generic_file_splice_read+0x360/0x5a0
 [<c1129e90>] ? page_cache_pipe_buf_release+0x20/0x20
 [<c105c56c>] ? enqueue_task_fair+0x1ac/0x620
 [<c105b584>] ? check_preempt_wakeup+0x144/0x220
 [<c1054b32>] ? check_preempt_curr+0x72/0x90
 [<c1051f5d>] ? ttwu_stat+0x2d/0x100
 [<c1055a6b>] ? get_parent_ip+0xb/0x40
 [<c1055a6b>] ? get_parent_ip+0xb/0x40
 [<c1670763>] ? sub_preempt_count+0x43/0xb0
 [<c102573e>] ? __kunmap_atomic+0x3e/0x90
 [<c13b22c0>] ? loop_make_request+0xf0/0xf0
 [<c13b19b9>] ? lo_splice_actor+0x89/0xc0
 [<c12c01e4>] ? security_file_permission+0x94/0xb0
 [<c112b83c>] generic_file_splice_read+0x6c/0xe0
 [<c112b7d0>] ? __generic_file_splice_read+0x5a0/0x5a0
 [<c112a201>] vfs_splice_to+0x61/0x80
 [<c112a2ca>] splice_direct_to_actor+0xaa/0x1d0
 [<c13b21b0>] ? loop_thread+0x510/0x510
 [<c13b1f43>] loop_thread+0x2a3/0x510
 [<c13b1ca0>] ? do_lo_send_write+0xe0/0xe0
 [<c1047103>] kthread+0x73/0x80
 [<c1047090>] ? kthread_freezable_should_stop+0x50/0x50
 [<c1673df6>] kernel_thread_helper+0x6/0x10

The fri2 has 1Gb of ram, hence the kernel sets 128Mb for the
allocation of vmalloc memory. The EMGD driver needs at least 256Mb of
vmalloc memory to work correctly, according to the EMGD documentation

This change passes the vmalloc allocation size to the kernel via the
vmalloc=256MB boot parameter.

Adjusting the kernel's vmalloc allocation size fixes the kernel errors.

This fixes bug:
[Yocto #3810]

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Acked-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

This replaces the old VA_FEATURES method for specifying codecs.

These VA codecs are included in the image when the image has 'hwcodec'
in the IMAGE_FEATURES variable.

One can disable the VA features by redefining the IMAGE_FEATURES
variable in local.conf.

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

This replaces the old VA_FEATURES method for specifying codecs.

These VA codecs are included in the image when the image has 'hwcodec'
in the IMAGE_FEATURES variable.

One can disable the VA features by redefining the IMAGE_FEATURES
variable in local.conf.

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

This replaces the old VA_FEATURES method for specifying codecs.

These VA codecs are included in the image when the image has 'hwcodec'
in the IMAGE_FEATURES variable.

One can disable the VA features by redefining the IMAGE_FEATURES
variable in local.conf.

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

This replaces the old VA_FEATURES method to specify the codecs.

These VA codecs are included in the image when image has 'hwcodec'
in the IMAGE_FEATURES.

And one can disable the VA features by redefining the IMAGE_FEATURES
var in the local.conf .

The gst-va-intel package pulls in emgd driver and x11 pieces, hence
it is moved to the XSERVERCODECS var.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

This replaces the old VA_FEATURES method to specify the codecs.

These VA codecs are included in the image when image has 'hwcodec'
in the IMAGE_FEATURES.

And one can disable the VA features by redefining the IMAGE_FEATURES
var in the local.conf .

The gst-va-intel package pulls in emgd driver and x11 pieces, hence
it is moved to the XSERVERCODECS var.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

This replaces the old VA_FEATURES method to specify the codecs.

These VA codecs are included in the image when image has 'hwcodec'
in the IMAGE_FEATURES.

And one can disable the VA features by redefining the IMAGE_FEATURES
var in the local.conf .

The gst-va-intel package pulls in emgd driver and x11 pieces, hence
it is moved to the XSERVERCODECS var.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

This replaces the old VA_FEATURES method to specify the codecs.

These VA codecs are included in the image when image has 'hwcodec'
in the IMAGE_FEATURES.

And one can disable the VA features by redefining the IMAGE_FEATURES
var in the local.conf .

The gst-va-intel package pulls in emgd driver and x11 pieces, hence
it is moved to the XSERVERCODECS var.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

The switchover to emgd neglected to add the user documentation needed
for whitelisting the video components.  This adds it and now matches
the other emgd-based BSPs.

Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com>

Submissions and questions about meta-intel BSPs should now go to the
new meta-intel mailing list.  Update MAINTAINERS with the new list.

Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

Signed-off-by: Andrei Gherzan <andrei@gherzan.ro>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

Signed-off-by: Andrei Gherzan <andrei@gherzan.ro>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

Move from 1.14 version to 1.16 version of the emgd driver for
graphics support.

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Acked-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

Move from 1.14 version to 1.16 version of the emgd driver for
graphics support.

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Acked-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

Create a new recipe for the latest emgd 1.16 version of the graphics
driver.

Also add the custom license files of the recipe to avoid build time
QA warnings.

There are strange characters in the license files, because they are
copied verbatim from the sources.

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Acked-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

We want to use emgd-1.14 & emgd-1.16 recipe interchangeably.
As the some of the generated package names include versions
which will be different in the emgd-1.16 recipe, this
version part is stripped from the package names so that it
will match with the emgd-1.16 recipe package names.

All the BSP configuration which were using these package names
also change accordingly to avoid any build issues.

Also with the 2 versions of emgd recipes make sure that BSPs
don't start using the emgd-1.16 version until they have been
validated to work by specifying the preferred emgd recipe
version for the relevant BSPs.

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Acked-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

There was a typo in specifying cfg/vesafb kernel feature. This commit
fixes it.

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Acked-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

As the commit to change kernel recipe SRC_URI to base on the
KSRC_linux_yocto_3_4_repo var did not go in the poky layer,
it leaves the SRC-URI definition incomplete here.
Fix the SRC_URI by specifying the git repository location
explicitly.

This fixes bug:
[YOCTO #3775]

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Acked-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

clean up unused variable.

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

The link has changed on the yoctoproject website, updating the URL
in the BSP layer READMEs.

This fixes the bug:
[Yocto #3730]

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

This is based on the gma500/gma600 kernel driver and the
modesetting X driver for poulsbo graphics.

The gma500 driver needs soft cursor in X.

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

The sys940x & sys940x-noemgd BSPs are still using v3.2 linux-yocto kernel.
This commit provides a v3.4 kernel recipe for these BSPs.

This addresses the bug/feature request:
[YOCTO #2268]

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

This change avoids addition of emgd source code in the noemgd bsp kernels.

This configures the fri2 BSP kernel with EMGD graphics driver
and fri2-noemgd BSP kernel with VESA graphics driver.

The VESA FB support is kept for both BSPs for boot time splash screen
support.

This addresses the bug/feature request:
[YOCTO #2268]

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

This change avoids addition of emgd source code in the noemgd bsp kernels.

This configures the emenlow BSP kernel with EMGD graphics driver.

The VESA FB support is kept for boot time splash screen support.

This addresses the bug/feature request:
[YOCTO #2268]

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

This change avoids addition of emgd source code in the noemgd bsp kernels.

This configures the crownbay BSP kernel with EMGD graphics driver
and crownbay-noemgd BSP kernel with just the VESA graphics driver.

The VESA FB support is kept for both BSPs for boot time splash screen
support.

This addresses the bug/feature request:
[YOCTO #2268]

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

This commit fixes this issue:
| In file included from gstvaapidisplay_glx.c:31:0:
| gstvaapiutils_glx.h:148:5: error: unknown type name 'PFNGLMULTITEXCOORD2FPROC'
| make[4]: *** [libgstvaapi_glx_0.10_la-gstvaapidisplay_glx.lo] Error 1

| ERROR: Function failed: do_compile (see /srv/home/nitin/build-test-bsps/build-sugarbay/tmp/work/x86_64-poky-linux/gstreamer-vaapi/0.3.8+gitAUTOINC+6ec4c2252a4aa706cd8631cb1083828485b9df9a-r0/temp/log.do_compile.73657

Found a solution to the issue here:
  https://bugs.freedesktop.org/show_bug.cgi?id=56516



Fixed the issue as per the comments on the above bug.

Fixes this bug:
[YOCTO #3722]

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

The file extension should be bbappend, sorry about that.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

Adjust bbappend to match changes in OE-Core.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

Fixes parsing errors which is appearing after this commit to
meta-openembedded

http://cgit.openembedded.org/meta-openembedded/commit/?id=3c21a46020bd0816579648f684c41dbd6333583e



This triggers
exception NameError: name 'base_contains' is not defined
without this change

Avoid the immediate expansion operator (:=) which caused base_contains
added in the previous commit to fail to expand at the time the
meta-intel layer configs are parsed and replace it with one of the
append operators (.= or +=) that allows for delayed variable expansion.

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Acked-by: Tom Zanussi <tom.zanussi@intel.com>
Signed-off-by: Darren Hart <dvhart@linux.intel.com>

Explanation from Terence Chiang:

"encountered a GFX issue while enabling HW video playback VC-1 and
H.264 simultaneously, the graphic driver report error with gfx hang on
Sandy Bridge platform. We worked with Intel Linux graphic team and
provided a patch"

Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>

The following EMGD subpackages provide video acceleration and have
dependencies on libva and libx11:

  emgd-driver-video
  emgd-gst-plugins-va-0.10.10
  emgd-gst-plugins-mixvideo-0.10.34

Include these in the VA_FEATURES variable only when the EMGD XSERVER
is included in the image.

For images such as core-image-lsb which do not include X components,
including any of these emgd subpackages will pull in unwanted X
components.  This fixes that problem.

Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Tom Zanussi <tom.zanussi@intel.com>