SIMD needs VSX with little endian to avoid the following build failure:
In file included from /nvmedata/autobuild/instance-12/output-1/build/jpeg-turbo-2.1.3/simd/powerpc/jccolor-altivec.c:25:
/nvmedata/autobuild/instance-12/output-1/build/jpeg-turbo-2.1.3/simd/powerpc/jccolext-altivec.c: In function 'jsimd_rgb_ycc_convert_altivec':
/nvmedata/autobuild/instance-12/output-1/build/jpeg-turbo-2.1.3/simd/powerpc/jsimd_altivec.h:93:26: warning: implicit declaration of function 'vec_vsx_ld'; did you mean 'vec_vsl'? [-Wimplicit-function-declaration]
93 | #define VEC_LD(a, b) vec_vsx_ld(a, b)
| ^~~~~~~~~~
Fixes:
- http://autobuild.buildroot.org/results/be6d5ad0cee4ee19eb25e595d44555a1af6e073b
Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
Signed-off-by: Arnout Vandecappelle <arnout@mind.be>
Currently, the jpeg-turbo tools are forcibly removed, but they can be
usefull (e.g. to test performance on the target, or to script batch
conversions).
Add an option to allow keeping them.
Signed-off-by: Clément Péron <peron.clem@gmail.com>
[yann.morin.1998@free.fr
- move the option out of the choice
- move the option to a Config.in.options
- drop the default 'n', which is ... the default
- use positive logic to define the hook
- move definition of the hook in the conditional block
- don't indent the hook assignment
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Since there is a variable definition between the definition of
BR2_PACKAGE_JPEG and the choice depending on it, the choice is
not indented bewlo the "jpeg support" prompt, like we like to
have:
[*] jpeg support
jpeg variant (jpeg-turbo) --->
Move the BR2_PACKAGE_JPEG right before the choice (really, move
BR2_PACKAGE_JPEG_SIMD_SUPPORT before BR2_PACKAGE_JPEG, but diff
finds the move of BR2_PACKAGE_JPEG is smaller to display):
[*] jpeg support
jpeg variant (jpeg-turbo) --->
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Similar to toolchains, we now offer a way for br2-external trees to
provide their libjpeg implementation, which gets included in the jpeg
choice.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Vadim Kochan <vadim4j@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Use easier-to-read layout; fix leading spaces.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
>From the 1.5-b1 release notes:
Added ARM 64-bit (ARMv8) NEON SIMD implementations of the commonly-used
compression algorithms (including the slow integer forward DCT and h2v2 &
h2v1 downsampling algorithms, which are not accelerated in the 32-bit NEON
implementation.) This speeds up the compression of full-color JPEGs by about
75% on average on a Cavium ThunderX processor and by about 2-2.5x on average
on Cortex-A53 and Cortex-A57 cores.
Add it unconditionally for all aarch64 cores, as neon support is required
for all "standard" ARMv8 implementations. If an ARMv8 implementation w/o
NEON ever shows up, then we will need to add a BR2_AARCH64_CPU_HAS_NEON and
handle it like ARM.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
>From the 1.5-b1 release notes:
Added full SIMD acceleration for PowerPC platforms using AltiVec VMX
(128-bit SIMD) instructions. Although the performance of libjpeg-turbo on
PowerPC was already good, due to the increased number of registers available
to the compiler vs. x86, it was still possible to speed up compression by
about 3-4x and decompression by about 2-2.5x (relative to libjpeg v6b)
through the use of AltiVec instructions.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In preperation for adding more architectures to the jpeg-turbo simd
handling, introduce a hidden BR2_PACKAGE_JPEG_SIMDS_SUPPORT symbol and use
that to enable simd support and default to jpeg-turbo, instead of open
coding it in both places.
While we are at it, reword the help text to be more useful.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Currently, the virtual package jpeg is a special virtual package,
as it offers a choice for its implementation, rather than letting the
user enable them manually.
In so doing, it defines its _HAS option as a 'def_bool y' rather than
letting each implementation 'select' it.
Since we are going to generate a list of virtual packages and their
providers, this defeats the heuristic used to find providers.
Coming with an alternate heuristic that also matches the jpeg
package is quite complex, so better and easier to make it look more
like any other virtual package.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Samuel Martin <s.martin49@gmail.com>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Since this package is implemented via a choice rather than the usual
separate-package providers, we only need BR2_PACKAGE_HAS_JPEG to be
always defined when the 'jpeg' package is selected.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Eric Le Bihan <eric.le.bihan.dev@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Cc: Mike Zick <minimod@morethan.org>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
For the underlying reasons, see previous patch, titled:
package/opengl/libegl: switch to package-defined providers
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Peter Korsgaard <jacmet@uclibc.org>
Cc: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
jpeg is a virtual package, but since it is listed in the dependencies
of other packages, it should obey to all the normal make rules for
packages. Notably, the jpeg-show-depends target is mandatory for the
graph-depends script to work.
Instead to implement such a make target manually, make jpeg a normal
generic-package, except that it doesn't have any source.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
'make source' fails because the host-jpeg-source target doesn't exist
anymore. Fix this by adding this target explicitly.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Introduce a virtual 'jpeg' package, which pulls in either libjpeg or
jpeg-turbo depending on a choice selection.
Rename jpeg package to libjpeg so we can reuse 'jpeg' for the virtual
package, making the change transparent to existing users and all the
packages using libjpeg.
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Signed-off-by: Will Newton <will.newton@imgtec.com>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Thanks to the pkgparentdir and pkgname functions, we can rewrite the
AUTOTARGETS macro in a way that avoids the need for each package to
repeat its name and the directory in which it is present.
[Peter: pkgdir->pkgparentdir]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
--program-prefix is now set in the auto-target infrastructure.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@anciens.enib.fr>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Now that we have libtool-2.2.x patch support, we can get rid of a bunch
of _LIBTOOL_PATH = NO, fixing (potential) cross link issues.
Notice: php not changed, as it uses a very old 1.5 version for the
embedded sqlite, where our buildroot-libtool-v1.5.patch doesn't apply.
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Buildroot as a whole is under GPL, so lets not expand each .mk file with
legal matter. The git history nicely shows the origin of each file, and
after the conversion to autotargets there basically isn't anything
left of the original file anymore.
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Now that <pkg>_INSTALL_TARGET_OPT always defaults to
'DESTDIR=$(TARGET_DIR) install', we can remove the
<pkg>_INSTALL_TARGET_OPT definition from a lot of packages.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
[Peter: has own libtool patch, use std staging install, fix target install]
Signed-off-by: Lionel Landwerlin <llandwerlin@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
We have been passing -q to ./configure when using 'make -s' for
packages using Makefile.autotools.in for some time. Do the same
for packages using autotools, but not using the
Makefile.autotools.in infrastructure, taking care to not do it
for packages with hand written configure scripts.
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
A C library will have been built by the toolchain makefiles, so there is no
need for packages to explicitly depend on uclibc.
Signed-off-by: Will Newton <will.newton@gmail.com>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
they should be configured with --prefix=/usr and we then need to use
make DESTDIR=$(STAGING_DIR) install to get things installed into the
staging directory. The current situation for many packages, which use
--prefix=$(STAGING_DIR) results in the staging_dir paths getting compiled
into the binary itself.
This also adds in a pile of libtool fixups. Between broken pkgconfig,
broken libtool handling, and broken --prefix settings, its a wonder
things have worked as well as they have up till now.
-Erik
Fabrice Fontaine