[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: libx86 for armhf (and armel)



Hi Anibal,

On 08/06/2012, Aníbal Monsalve Salazar <anibal@debian.org> wrote:
> On Mon, Feb 20, 2012 at 02:47:12AM +0200, Philippe De Swert wrote:
>>Hi,
>>
>>Does anybody know how to test if libx86 works on armhf/armel?
>>As I have used a slightly different approach to get this to compile.
>>Turns out the offending assembly in
>>http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=597000 is just a
>>macro which generates the bog standard inx (where x is b, w or l) and
>>similar for outx functions in assembly.
>>Looking at the libc code for x86 it is indentical and it seems to
>>offer some similar functionality for arm, but I could not immediately
>>locate the asm for it.
>>Anyhow going on the hunch that such basic functionality probably works
>>and the inx and outx functions are nicely defined in <sys/io.h> I just
>>tried some ugly refdefinition when we use the x86emu to compile (as is
>>the case on non-x86 platforms) (see patch attached). And it built fine
>>after that.
>>I am actually wondering if there is a reason to use asm instructions
>>there... so that could actually been cleaned up totally.
>>
>>So does anybody know how to test this thing? And confirm if my hunch
>>actually delivers a working solution?
>>
>>Thanks,
>>
>>Philippe
>
> [The attached file is Philippe's patch.]
>
> Please read http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=570676#29

Well yeah that is exactly what I say. Except that libc actually has
calls to emulate this inx and outx calls on arm. So the patch should
work. I just don't know how to test it to be sure it works correctly.
So as you seem to be the libx86 maintainer I have a patch that will
most likely work and does allow us to have libx86 on arm(el/hf). If
you could give me an idea how to test it that would be very welcome. I
can also give you access to an armhf machine if you would like to do
it yourself. Well worst case I guess I can ask the libx86 guys I
guess.

Cheers,

Philippe


Reply to: