-Wl,--fix-cortex-a8 is a workaround for an Erratum in Cortex-A8
processors. It slightly increases code size and decreases performance,
and there's no point in using it on non-A8 CPUs.
Instead of forcing it unconditionally, use it when targeting
Cortex-A8 or generic armv7-a (which might or might not be A8).
Change-Id: Ifa59765d380445237edccfe5440a67b3ba1e459a
Signed-off-by: Bernhard Rosenkränzer <Bernhard.Rosenkranzer@linaro.org>
Otherwise wrong set of memory/string functions is chosen when building atom
image (in Bionic we have libc/arch-x86/atom/atom.mk). Plus this naming is
consistent with other x86 architectures.
Change-Id: I21e899534e7ce10530474a22ceba770422b39d8d
Signed-off-by: Alexander Ivchenko <alexander.ivchenko@intel.com>
Atomic functions used in external/libcxx/include/atomic when compiled with Clang
will require intrinsic functions exist only for prescott or newer CPUs.
BUG: 17530542
Change-Id: I0c9660ed2ffa75b940981eb8165d88934b39aec5
LPAE indicates better instructions can be used when atomicity guarantees are
needed. However, LPAE's presence isn't advertised by clang/GCC. We fake an
ARM feature to advertise its presence on architectures where it is.
Also, add a TODO documenting that cortex-a15 is not the correct CPU variant
for krait.
Change-Id: I02a1248025c32d94eca0bc8a249dc524f1ac9c36
For ndk docs change, please refer to:
https://android-review.googlesource.com/#/c/110100/
Change-Id: I8428e7a979eb02441066aeeee43ce693d4d0dc8d
Signed-off-by: Alexander Ivchenko <alexander.ivchenko@intel.com>
Delete x86_64-atom.mk as we don't support 64-bit on old Atom.
Change-Id: I0b9ab61cd9b840f32c30059cb3ba9704c733c42a
Signed-off-by: Varvara Rainchik <varvara.rainchik@intel.com>
Add mips64r6 target and corresponding mips32r6 target.
Defaults remain as mips64r2 and mips32r2.
Apply -FP64A codegen subsetting to mips32r6 only.
Access FR=0 odd-numbered 32-bit float regs only via
double-prec even-numbered regs, not by single-prec ops.
Change-Id: I1740a6c658304b6c41242be58d68753e6f171658
Use 4.9 mips64el toolchain for both 64- and 32-bit builds.
Tell ld when 32-bit links are required.
Override 4.9's changed defaults for mips floating point
register use, to get same assembler rules as 4.8 and earlier.
Also: drop unused soft-fp build targets, cleanout redundant
compiler options, and remove extraneous Android.mk file.
Change-Id: I86f1075266349edb2b08a7709b9f5472d8cfda32
This is used for Baytrail targets.
Change-Id: I5a2fa6dbb8217a326ee09f5ea434885718ab3f0c
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Signed-off-by: Pavel Chupin <pavel.v.chupin@intel.com>
Signed-off-by: Fengwei Yin <fengwei.yin@intel.com>
Users of ARCH_ARM_* defines don't care about first vs. second arch,
set ARCH_ARM_* regardless of which arch is arm.
Change-Id: I2ae83ec5c3f839ff91a0e352c95d76ec2cbd5dc5
This is the first step to build 32-bit libraries in a 64-bit product.
It will work like this:
1) In the product's BoardConfig.mk, define:
TARGET_2ND_ARCH, TARGET_2ND_ARCH_VARIANT, TARGET_2ND_CPU_VARIANT.
The build system uses those variables to set up an additional compiler
environment for the second arch.
2) When parsing Android.mks, the build system sets up rules to build a
module for both the 1st arch and the 2nd arch, unless it's explicitly
asked to skip so.
Android.mk will be adapted if there is additional rule of generating
source files.
The build system will accept arch-specific LOCAL_ variables, such as
LOCAL_CFLAGS_arm, LOCAL_CFLAGS_armv7-a-neon, LOCAL_CFLAGS_cortex-a15,
LOCAL_CFLAGS_aarch64 etc. Modules use such variables to set up build for
various archs at the same time.
3) Install binary of the 2nd arch by adding "<module_name>:32" to
PRODUCT_PACKAGES. All 2nd-arch libraries linked in by "<module_name>:32"
will be installed automatically.
Bug: 11654773
Change-Id: I2df63cd5463a07bf5358bee2a109f8fb9590fe30
Conflicts:
core/combo/TARGET_linux-arm.mk
ARCH_X86_HAVE_MMX, ARCH_X86_HAVE_SSE, ARCH_X86_HAVE_SSE2,
and ARCH_X86_HAVE_SSE3 were all always on. There are no longer any makefiles
or code that are conditional on any of these, so let's just remove them
rather than encourage anyone to mess with knobs that don't work.
Change-Id: I5ee095e8041eecff4554ad4801894fbfca69093f
Add x86_64 Android builds. Compiler is expected to be able to understand
-m64 code generation option.
Change-Id: I99e7337c5a5766afc5e528a481bd21631ff44dd5
Signed-off-by: Pavel Chupin <pavel.v.chupin@intel.com>
Author: Negreanu Marius Adrian <adrian.m.negreanu@intel.com>
Author: Andrew Boie <andrew.p.boie@intel.com>
Author: Daniel Leung <daniel.leung@intel.com>
Currently, x86 target only has generic i686 and x86-atom
as arch variants. This patch adds the ability to have
more than two arch variants. Defining a new arch variant
is similiar to ARM targets, by adding a new file in
core/combo/arch/x86. These files also define what
capabilities the targeting CPU has (e.g. having SSE2,
SSE3, etc.).
We define arch variants for Sandy Bridge, Ivy Bridge,
Haswell; upcoming arches can be easily added to this
set with future patches.
Change-Id: Iafbce10d205e860738db4a216ff603f9a84d7311
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
synci does not provide coherency between CPU's on this device
Change-Id: I10e73fa49859e55d018884c6682b5a00b887e0a1
Signed-off-by: Chris Dearman <chris.dearman@imgtec.com>
* commit '7ce7473f866fc5356291f38a3437a071ea5fd00d':
For the current MIPS compiler __builtin___clear_cache() generates synci instruction only with -msynci option So, add -msynci to all mips32r2 makefiles. Also add msynci to the list flags not recognized by clang.
