| Age | Commit message (Collapse) | Author |
|---|
|
This reverts commit 5a739a95bcb202a5f2fc0d928da3669585aba64a.
In meta-openembedded: "d95712f mozjs: fix armeb builds"
Change-Id: I124a8314b2904c39d1586d97f3381c7bc482b384
Signed-off-by: Anders Roxell <anders.roxell@linaro.org>
|
|
Change-Id: Ic0bc07c6ef378486fd8752f98de7b43798055982
Signed-off-by: Koen Kooi <koen.kooi@linaro.org>
|
|
Change-Id: I5a2c903d9f9f0c0475804ba369fc9f1738ce4213
Signed-off-by: Koen Kooi <koen.kooi@linaro.org>
|
|
alignment.sh is only really needed on armv5 anyway
Change-Id: I6488b37fcb05e2f8854d362f9178ef7bf56fd9d2
Signed-off-by: Koen Kooi <koen.kooi@linaro.org>
|
|
OE-core has dropped 2.7.3 a while ago and no bugreports came in, so drop this bbappend.
Change-Id: I9cca1a40043c572d564cb6f621912369dff57e22
|
|
OE-core has the armeb config-uri in the recipe.
Change-Id: I8570e2948e7737a7c68ea816bfaaa937740b4a60
Signed-off-by: Koen Kooi <koen.kooi@linaro.org>
|
|
Change-Id: I8295db1b1fc64af03f8f262a2a36fd5a5faf5b9d
Signed-off-by: Koen Kooi <koen.kooi@linaro.org>
|
|
Change-Id: Ief1aa78fe443cf6fdef9716656df4badb21000a5
Signed-off-by: Fathi Boudra <fathi.boudra@linaro.org>
|
|
Change-Id: I8e3c628ed97c5c1310f3d84ca06b4f2c75c72132
Signed-off-by: Gary S. Robertson <gary.robertson@linaro.org>
|
|
Change-Id: Ice7be4f5acb6a6e0c55dda53a0a22c2675df93d9
Signed-off-by: Koen Kooi <koen.kooi@linaro.org>
|
|
* patch is now merged upstream
* oe-core recipe is updated to 2.0.6 release (and contains the patch)
Change-Id: I6fc058fd4c3a754f99fb4ea16bc4db6649aca0c1
Signed-off-by: Fathi Boudra <fathi.boudra@linaro.org>
|
|
Change-Id: Ic0ae08a66e91c7855c521d95ea831183b5ef73a8
Signed-off-by: Taras Kondratiuk <taras.kondratiuk@linaro.org>
|
|
Change-Id: I73b4ea9ed061c2c728b21709fbb11e2a6e0eedea
Signed-off-by: Koen Kooi <koen.kooi@linaro.org>
|
|
http://cgit.openembedded.org/meta-openembedded/commit/?id=9fede488dc6f1320f6000be93953a887ade52eb7
Change-Id: I8826d62ed80213e6ad62590f3e2005d70d0bfe23
Signed-off-by: Fathi Boudra <fathi.boudra@linaro.org>
|
|
Change-Id: Ie71781667200040ba512b8c40aaae465ab495d94
Signed-off-by: Victor Kamensky <victor.kamensky@linaro.org>
Signed-off-by: Koen Kooi <koen.kooi@linaro.org>
Acked-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Acked-by: Fathi Boudra <fathi.boudra@linaro.org>
|
|
Change-Id: I6241741de826b17d4b438730ce9a021f88c7ff6b
Signed-off-by: Anders Roxell <anders.roxell@linaro.org>
|
|
|
|
Since python source base has its own copy of libffi, it fixes the issue
here. Note similar fix is committed into separate libffi library.
cffi code creates trampoline for python callback function, which is
supposed to be called from another C function and it does not take into
account that in ARMv7a in big endian mode instructions are still in
little endian. Since written by big endian code they have to be
byteswapped. Proposed fix uses byteswap codes for instructions.
Here is python code snippet that will fail when libffi is used by
ctypes in python code (func1 and func2 are C functions):
from ctypes import *
mylib = CDLL("libmytest.so")
func1 = mylib.func1
print func1(1, 2)
CBFUNC = CFUNCTYPE(c_int, c_int, c_int)
def myfunc2(a, b):
return a * a + b * b
cb_func = CBFUNC(myfunc2)
func2 = mylib.func2
print func2(cb_func, 1, 2)
code will produce illegal instruction when func2 will call myfunc2
through trampoline
Signed-off-by: Victor Kamensky <victor.kamensky@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
|
|
cffi code creates trampoline for python callback function, which is
supposed to be called from another C function and it does not take into
account that in ARMv7a in big endian mode instructions are still in
little endian. Since written by big endian code they have to be
byteswapped. Proposed fix uses byteswap codes for instructions.
Here is python code snippet that will fail when libffi is used by
ctypes in python code (func1 and func2 are C functions):
from ctypes import *
mylib = CDLL("libmytest.so")
func1 = mylib.func1
print func1(1, 2)
CBFUNC = CFUNCTYPE(c_int, c_int, c_int)
def myfunc2(a, b):
return a * a + b * b
cb_func = CBFUNC(myfunc2)
func2 = mylib.func2
print func2(cb_func, 1, 2)
code will produce illegal instruction when func2 will call myfunc2
through trampoline
Signed-off-by: Victor Kamensky <victor.kamensky@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
|
|
on ARM V7 operating in big endian mode strace does not work:
root@genericarmv7ab:~# strace ls
pid 1356 unknown syscall trap 0x000000ef
it happens because ARM V7 when runs as big endian operates in be8 mode,
where instruction are still in little endian form. Strace get_scno reads
instructions and matches it to certain pattern, but in armeb case it needs
to byteswap it before that.
Signed-off-by: Victor Kamensky <victor.kamensky@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
|
|
dds the same patches into SRC_URI_append_armeb as
SRC_URI_append_arm does.
Signed-off-by: Victor Kamensky <victor.kamensky@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
|
|
Adds the same patches into SRC_URI_append_armeb as
SRC_URI_append_arm does.
Signed-off-by: Victor Kamensky <victor.kamensky@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
|
|
Create meta-linaro/meta-bigendian meta layer that holds oe changes
to build ARM big endian images
Signed-off-by: Victor Kamensky <victor.kamensky@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
|
