[Interest] Qt5 cross compilation issues on Tegra3

Ramakanthreddy_Kesireddy Ramakanthreddy_Kesireddy at mahindrasatyam.com
Fri Feb 22 12:42:00 CET 2013


Hi,

Actually as neon support is enabled for Tegra3 targetfs, -mfloat-abi=soft is disabling the neon.
Please suggest the changes we need to make in qmake.conf as attached.

Br,
Ramakanth
________________________________________
From: interest-bounces+ramakanthreddy_kesireddy=mahindrasatyam.com at qt-project.org [interest-bounces+ramakanthreddy_kesireddy=mahindrasatyam.com at qt-project.org] on behalf of Donald Carr [sirspudd at gmail.com]
Sent: Friday, February 22, 2013 7:06 AM
To: Thiago Macieira
Cc: interest at qt-project.org
Subject: Re: [Interest] Qt5 cross compilation issues on Tegra3

Yuck, I suspect I might be somewhat to blame. There is so much wrong,
I might be causing some of it.

If you create a "device" (TM) mkspec as evident in the device subdir,
it automatically includes a header and footer pri file, which try to
do the right thing but in a braindead misguided fashion.

One of the things that the footer does is automatically append
-mfloat-abi=softfp unless you explicitly state that you are using
hardfp support. My justification for this stupidity is below, if you
care.

This is limited to mkspecs in the device folder, and this was meant to
ease optimal usage/compilation of Qt since:

1) Our documentation doesn't cut it (cross compilation)
2) Our customers are normally complete dolts, even the chipset dudes
who appear to love living in complete ignorance of modern Linux
stacks, toolchain progress and the impact of compiler flags on
generated code.

Since we never actually got the device crap to a point where it was
widely adopted, understood and used, this lingers as a semi-functional
documentational sore point.

I have meant to address this oversight for absolute ages.

Cheerio,
Donald


========================================

This was introduced due to the us (possibly just me) deciding that we
always wanted -mfloat-abi=softfp _or_ -mfloat-abi=hard, and never
-mfloat-abi=soft which is the default for most toolchains I have
encountered as part of BSP packages.

On devices like the Raspberry Pi -mfloat-abi=softfp made a hell of a
difference over -mfloat-abi=soft, and at a point where the hardfp
enabled build did not exist.



On Thu, Feb 21, 2013 at 8:13 AM, Thiago Macieira
<thiago.macieira at intel.com> wrote:
> On quinta-feira, 21 de fevereiro de 2013 10.28.37, Ramakanthreddy_Kesireddy
> wrote:
>> Please let me know how to resolve the same.
>
> One of the errors is clear:
>
> xcb.cpp:42:21: fatal error: xcb/xcb.h: No such file or directory
>
> You need to install the necessary development packages.
>
> The other error is that your toolchain is broken.
>
> Your compiler flags include:
>         -mhard-float -mfloat-abi=softfp -mfpu=vfpv3-d16 -mfloat-abi=softfp
>
> Your linker flags don't.
>
> Why are you compiling new software using "softfp" ?
>
> --
> Thiago Macieira - thiago.macieira (AT) intel.com
>   Software Architect - Intel Open Source Technology Center
>
> _______________________________________________
> Interest mailing list
> Interest at qt-project.org
> http://lists.qt-project.org/mailman/listinfo/interest
>
_______________________________________________
Interest mailing list
Interest at qt-project.org
http://lists.qt-project.org/mailman/listinfo/interest

________________________________

DISCLAIMER:
This email (including any attachments) is intended for the sole use of the intended recipient/s and may contain material that is CONFIDENTIAL AND PRIVATE COMPANY INFORMATION. Any review or reliance by others or copying or distribution or forwarding of any or all of the contents in this message is STRICTLY PROHIBITED. If you are not the intended recipient, please contact the sender by email and delete all copies; your cooperation in this regard is appreciated.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: qmake.conf
Type: application/octet-stream
Size: 1270 bytes
Desc: qmake.conf
URL: <http://lists.qt-project.org/pipermail/interest/attachments/20130222/0cbaab6b/attachment.obj>


More information about the Interest mailing list