Support cross compiling for armhf variant ARM/linux.

Support cross compiling for armhf variant ARM/linux.

armhf (hardware float) is the ubuntu default these 
days so should be used over armel in most cases.
Also, the chromium ARM/linux test hardware is 
running armhf.

This CL does not change the default or remove armel
support. That is left for a future CL.

BUG=308256
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=231487

[Sam Clegg, 2013-10-28 17:20:09]

https://codereview.chromium.org/27197011/diff/52001/build/common.gypi
File build/common.gypi (right):

https://codereview.chromium.org/27197011/diff/52001/build/common.gypi#newcode...
build/common.gypi:1770: 'arm_float_abi%': 'hard',
Really I would have liked to remove this line completely and let the
compiler's default abi be used but removing cased problems at some
point in the build with '' being and invalid value for --mfloat-abi.

I could investigate further...

[ghost stip (do not use), 2013-10-28 18:17:02]

https://chromiumcodereview.appspot.com/27197011/diff/52001/build/common.gypi
File build/common.gypi (right):

https://chromiumcodereview.appspot.com/27197011/diff/52001/build/common.gypi#...
build/common.gypi:1770: 'arm_float_abi%': 'hard',
On 2013/10/28 17:20:09, Sam Clegg wrote:
> Really I would have liked to remove this line completely and let the
> compiler's default abi be used but removing cased problems at some
> point in the build with '' being and invalid value for --mfloat-abi.
> 
> I could investigate further...

Looking at http://gcc.gnu.org/onlinedocs/gcc/ARM-Options.html I don't see a
'system' option for -mfloat-abi so I guess we'd have to eliminate the option
altogether to get the system type. I'm OK with hardcoding it here

https://chromiumcodereview.appspot.com/27197011/diff/52001/build/linux/instal...
File build/linux/install-arm-sysroot.py (right):

https://chromiumcodereview.appspot.com/27197011/diff/52001/build/linux/instal...
build/linux/install-arm-sysroot.py:25:
nativeclient-archive2/toolchain/$NACL_REV/sysroot-arm-trusted.tgz
why are we changing the name of the tgz?

[ghost stip (do not use), 2013-10-28 18:19:43]

you know what the deal is with the '[library] uses VFP register arguments,
output does not' that show up in the tryjob? do we have to update the
cross-compilation libs on the bot?

[Sam Clegg, 2013-10-28 18:21:01]

https://chromiumcodereview.appspot.com/27197011/diff/52001/build/linux/instal...
File build/linux/install-arm-sysroot.py (right):

https://chromiumcodereview.appspot.com/27197011/diff/52001/build/linux/instal...
build/linux/install-arm-sysroot.py:25:
nativeclient-archive2/toolchain/$NACL_REV/sysroot-arm-trusted.tgz
On 2013/10/28 18:17:02, stip wrote:
> why are we changing the name of the tgz?

I changed this on the nacl side since the old name seemed misleading.

We currently re-use the sysroot images from NaCl in the chrome build but perhaps
chrome should have its own sysroot image like it does for the x86 sysroot image:

https://code.google.com/p/chromium/codesearch#chromium/src/chrome/installer/l...

[Sam Clegg, 2013-10-28 18:23:18]

On 2013/10/28 18:19:43, stip wrote:
> you know what the deal is with the '[library] uses VFP register arguments,
> output does not' that show up in the tryjob? do we have to update the
> cross-compilation libs on the bot?

We need to land the trybot config change before those will go away:

https://codereview.chromium.org/49233002/

[ghost stip (do not use), 2013-10-28 18:23:39]

ah, do the changes in https://chromiumcodereview.appspot.com/49233002/ fix
those?

[Lei Zhang, 2013-10-28 21:23:17]

https://codereview.chromium.org/27197011/diff/222001/build/common.gypi
File build/common.gypi (right):

https://codereview.chromium.org/27197011/diff/222001/build/common.gypi#newcod...
build/common.gypi:1773: # (normally arm-linux-gnueabi-gcc or
arm-linux-gnueabihf-gcc)
Mention the permissible values are soft, softfp and hard?

https://codereview.chromium.org/27197011/diff/222001/build/install-build-deps.sh
File build/install-build-deps.sh (right):

https://codereview.chromium.org/27197011/diff/222001/build/install-build-deps...
build/install-build-deps.sh:144: arm_list+=" $armel_list"
We generally append with foo="$foo bar" here.

https://codereview.chromium.org/27197011/diff/222001/build/install-build-deps...
build/install-build-deps.sh:145: echo $arm_list
debugging statement?

[Sam Clegg, 2013-10-28 21:26:18]

Support both armel and armhf until the transition is
complete.

https://codereview.chromium.org/27197011/diff/222001/build/install-build-deps.sh
File build/install-build-deps.sh (right):

https://codereview.chromium.org/27197011/diff/222001/build/install-build-deps...
build/install-build-deps.sh:144: arm_list+=" $armel_list"
On 2013/10/28 21:23:17, Lei Zhang wrote:
> We generally append with foo="$foo bar" here.

Done.

https://codereview.chromium.org/27197011/diff/222001/build/install-build-deps...
build/install-build-deps.sh:145: echo $arm_list
On 2013/10/28 21:23:17, Lei Zhang wrote:
> debugging statement?

Done.

[Lei Zhang, 2013-10-28 21:31:50]

build/common.gypi in patch set 4 no longer reflects the CL description. Are you
going to make the switch in this CL or use it to set up things for the
transition?

[Sam Clegg, 2013-10-28 21:58:13]

On 2013/10/28 21:31:50, Lei Zhang wrote:
> build/common.gypi in patch set 4 no longer reflects the CL description. Are
you
> going to make the switch in this CL or use it to set up things for the
> transition?

Thanks.

Updated CL.

[Lei Zhang, 2013-10-28 22:05:48]

lgtm

[commit-bot: I haz the power, 2013-10-28 23:41:28]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/sbc@chromium.org/27197011/422001

[commit-bot: I haz the power, 2013-10-29 02:38:48]

Change committed as 231487