Using CORTEX_ID variable reference in the tuning overrides did not work.
This reverts those changes, and adds a tuning file for the cortex-a5.
Revert "tune-cortexa5.inc: Add tune file for cortex-a5"
Revert "tune-cortexa.inc: create a common include for cortex-a armv7a tuning"
(From OE-Core rev: 74158c2e99c6d8631800ae80025d1cc9f19336d2)
Signed-off-by: Andy Voltz <andy.voltz@timesys.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The tuning files for the cortex-a* processors are mostly identical for
the A7,A8,A9,A15 processors. Rework these files to use a CORTEX_ID
variable to setup the tuning for each specific processor.
(From OE-Core rev: 3e4f4a1cf07ff7cf4c71566492385f8fbf581789)
Signed-off-by: Andy Voltz <andy.voltz@timesys.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* number of TUNE_CCARGS conditionals is important if we add
extra space with each one in "else" branch
I'm building for 2 MACHINEs one is cortexa9, second is cortexa8
few months ago we added TUNE_CCARGS[vardepvalue] in bitbake.conf
http://git.openembedded.org/openembedded-core/commit/?id=03f1e34ea3ce80931e9c3cd2ab22824f28a7233b
which fixed some cases (like mentioned tune-xscale and tune-arm926ejs)
where both had unused TUNE_CCARGS when common DEFAULTTUNE was used.
with cortexa[89] it's different, because cortexa9 has one extra TUNE_CCARGS
TUNE_CCARGS += "${@bb.utils.contains("TUNE_FEATURES", "cortexa9", "-mtune=cortex-a9", "", d)}"
which adds extra *space* even when not used because of '+=' and as result:
$ bitbake-diffsigs tmp-eglibc/sstate-diff/1366797730/*/armv7*/adapterbase/*do_configure*
basehash changed from f986789fb8fb3579ed6a3492cc8a8d10 to c851b5f838d945ee13072e9ad6725dca
Variable TUNE_CCARGS value changed from
' -march=armv7-a -mthumb-interwork -mfloat-abi=softfp -mfpu=neon '
to
' -march=armv7-a -mthumb-interwork -mfloat-abi=softfp -mfpu=neon '
Hash for dependent task gcc-runtime_4.7.bb.do_populate_sysroot changed from bdeabf7a86958b9110b566344b7916de to 2be5618e6bc8c57ec9db5659bf217915
Hash for dependent task eglibc_2.17.bb.do_populate_sysroot changed from b4f40fc62dde684acd0a574532a55360 to 97fcb426603d4a1c1099c0504d2ebf7d
Hash for dependent task glib-2.0_2.34.3.bb.do_populate_sysroot changed from fd2f90b83098c34e88d649d70f6ea4f5 to ebd740bb94ea3eb0a914efda6fc82c4a
(From OE-Core rev: b7430ff83760ac29079d20dc7c62f498a0a9d55d)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
OVERRIDES reads from left to right, least to most specific. We were
appending to MACHINEOVERRIDES when we should have been prepending so
the ordering of qemuall verses qemuxxx was incorrect, as was the x86
override and several of the arm overrides. This patch is a batch cleanup
of the various issues to correct the order from least to most specific.
The include order does matter and we needed to tweak some of that in this
patch too.
[YOCTO #4090]
(From OE-Core rev: bdc1b214431c9c93a929b547b9a61e7b87fbd366)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This work was made by Victor Enriquez and then modified by Denis Carikli
who was helped by Mark Hatle comments. And in the end modified by Martin
Jansa to support different ARMPKGARCH and removed explicit -novfp suffix.
The changes are for adding support to armv6-novfp, for building binaries
for armv6 machines without vfp, for example the htc dream.
(From OE-Core rev: 0733e2f8ad82b426c8c40ef753adb9431fa3c359)
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@no-log.org>
Signed-off-by: Víctor Enríquez <victor.quicksilver@gmail.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* without this patch it does apply --fix-v4bx not only to armv4, but
also all higher (because they also have armv4 in TUNE_FEATURES)
* it causes SIGILL on armv4t
http://lists.linuxtogo.org/pipermail/openembedded-devel/2012-November/042298.html
* someone please test on armv4 device (I tested only bitbake -e output
that it's correctly applied with DEFAULTTUNE == armv4
* maybe we can should fix this in binutils instead (both 2.22 and 2.23
are affected)
(From OE-Core rev: efe03fc00fc051bede69ced6643a8f25d02eabde)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* that we always use TUNE_FEATURES_tune-arm* variable and add only one TUNE_FEATURE to it
* for bigendian always use littleendian counterpart and append bigendian TUNE_FEATURE
(From OE-Core rev: 1bc205f895c8143e0bde3c4ba0e699cc0b2f0de8)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* bigendian should not include little endian PACKAGE_ARCHS
(From OE-Core rev: 42e18249b02280de28fb7159b11e3c7c78a6cb03)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* tune-foo is not valid override, for it to work I had to add
ARMPKGARCH = "${ARMPKGARCH_tune-${DEFAULTTUNE}}"
but that doesn't work without value defined for every supported
DEFAULTTUNE value, otherwise it's expanded like this
TUNE_PKGARCH (${ARMPKGARCH_tune-armv5te}te).
(From OE-Core rev: 31e4f2dee990ee7f5d7491b65565e71d7d580209)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We have been adding this option to paper over a bug in old toolchain
http://hardwarebug.org/2008/11/28/codesourcery-fails-again/
e.g. is one but these have been weeded out. Therefore let gcc
take the default vectorization optimizations
(From OE-Core rev: e4336ab56db1e07a7f3dc08d3a4de3593b0fad22)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This cleans up and/or corrects a few values from machine includes
for consistency with future toolchain sanity checks, and also adds
the TUNEVALID and TUNECONFLICTS to documentation.conf.
(From OE-Core rev: 6ffe53c721a80cf156b44f59b564f2e899c6af50)
Signed-off-by: Peter Seebach <peter.seebach@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Cleanup the ARM tunings to match the new tunings README file.
The ARM tunings define TUNE_PKGARCH in a way that only one main
arm architecture, i.e. armv6, may be defined at the same time. We
may have to revise these settings in the future, as well as figure
out a way to better differentiate various optimize tunings in the
package arch. (This was not done, to preserve existing behavior!)
Fix a number of minor issues w/ the armv5 tunings where DSP variants
were referenced but not defined.
Fix incorrect armv7 entries in armv7a.
Fix PACKAGE_EXTRA_ARCHS definitions inside of tune-cortexm3 and tune-cortexr4.
(From OE-Core rev: 0e71abea5458122188d5eddef2c17147f61ff895)
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
armv7 is least common denominator of armv7-a
armv7-m and armv7-r and armv7-m does not support
ARM instructions but only thumb2 instruction set
which means armv7 when chosen will complain if
code is compiled in arm mode which is default
in OE if not specified other wise
if we chose this tuning errors like below pop up
error: target CPU does not support ARM mode
This tuning seems theoretical and base tune
for armv7 would be one of armv7-a, armv7-m or
armv7-r
(From OE-Core rev: 75b8adbc042e0f65fb1286bc550d02becd3b6aea)
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Using "1" with getVar is bad coding style and "True" is preferred.
This patch is a sed over the meta directory of the form:
sed \
-e 's:\(\.getVar([^,()]*, \)1 *):\1True):g' \
-e 's:\(\.getVarFlag([^,()]*, [^,()]*, \)1 *):\1True):g' \
-i `grep -ril getVar *`
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Recent versions of the GCC reject the -mno-thumb option. In order to prevent
the compiler from generating code for the Thumb instruction set the -marm
switch should be used instead. For details see GNU bug #47930.
(From OE-Core rev: 72dc73f5a647ccd38145fd888c109a144f202963)
Signed-off-by: Ken Werner <ken.werner@linaro.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* motivated by this NAK
http://patchwork.openembedded.org/patch/15777/
and today's discussion on #yocto I hope it's worth it to send this RFC
(From OE-Core rev: e3e1fef27345e2ea923b76b1e6bcb9cd5572cec6)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
A closing quote was missing for an AVAILTUNES append operation, add it.
(From OE-Core rev: 7d46901840795638ff184a43e65299446fbd0b4e)
Signed-off-by: Darren Hart <dvhart@linux.intel.com>
CC: Jason Kridner <jkridner@beagleboard.org>
CC: Koen Kooi <koen@dominion.thruhere.net>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The previous commit to this file meant thumb was always being turned on
even when TUNE_FEATURES did not contain "thumb". This is clearly wrong
and this patch corrects this so thumb options are no longer specified
in that case.
(From OE-Core rev: 4b5e8074f8aca59b09421db464ce652e84f898f2)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
PACKAGE_EXTRA_ARCHS_tune-armv5eb needs to be defined in terms of
the non-e with the same endianness, i.e. PACKAGE_EXTRA_ARCHS_tune-armv5b
not PACKAGE_EXTRA_ARCHS_tune-armv5, otherwise PACKAGE_EXTRA_ARCHS will
end up containing a semi-random mixture of endiannesses and disaster
will ensue. Likewise for the vfp and armv6 variants.
This is all a bit confusing because TUNE_FEATURES is done the opposite
way around, i.e. TUNE_FEATURES_tune-armv5eb is derived by taking the
armv5e version and adding bigendian. But fixing that is probably
a subject for a separate patch.
(From OE-Core rev: 391c0102a81455c76244d13b6878e3a76cca65dc)
Signed-off-by: Phil Blundell <philb@gnu.org>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
These changes revolve around the idea of tune features. These are represented by
'flag' strings that are included in the TUNE_FEATURES variable.
Any string included in TUNE_FEATURES should also add a TUNEVALID[<name>] entry so
we can know which flags are available in TUNE_FEATURES and have documentation about
what the flags do. We will add sanity code to error if flags are listed in
TUNE_FEATURES but are not documented in TUNEVALID.
A given tune configuration will want to define one or more predetermined sets of
_FEATURE flag lists. These are defined in the form TUNE_FEATURES_tune-<name>.
For defined tune configuation, <name> should be added to the AVAILTUNE list so that
we can determine what tune configurations are available. Flags cannot be used in this
case as with TUNEVALID since its useful to be able to build up tune lists from other
TUNE_FEATURES_tune-yyy options.
A given tune configuration may also define PACKAGE_EXTRA_ARCHS_tune-<name> and
BASE_LIB_tune-<name> to control the multilib location. All options can be overridden
by the distro or local user configuration.
(From OE-Core rev: 5f9d56bd64997b93ed7e46c117851002a0556654)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Andy Voltz
Martin Jansa
Richard Purdie
Víctor Enríquez
Khem Raj
Peter Seebach
Mark Hatle
Ken Werner
Darren Hart
Koen Kooi
Phil Blundell