Hi,

On Wed, 22 Apr 2009 12:21:01 +0300
Eduard - Gabriel Munteanu <[email protected]> wrote:

> Some code relating to x86 and x86-64 differs only in regards to register
> width. This change makes it possible to specify registers by word-size
> independent names, thus allowing for generic code while keeping the
> compatibilities visible.

I do not love this change.

Here's part of the relevant IRC log:

<penberg> ahuillet: to share code with 32-bit and 64-bit
<penberg> anything wrong with that?
<ahuillet> yes, what is the point of it ?
<ahuillet> in what case do we not care about the size of our operands ?
<penberg> ahuillet: use-def
<ahuillet> for use-def it is only useful if we assume that the LIR for both 
architectures are going
to be the same 
<penberg> the same or at least very similar, yes.
<ahuillet> well for now I don't feel good about this change because it's very 
confusing to have
registers called XAX 
<ahuillet> couldn't we instead have REG_EAX, REG_RAX = REG_EAX in the enum ?
<penberg> ahuillet: hmms, REG_EAX is a real register on x86-64.
<penberg> that's why XAX makes sense to me
<ahuillet> and I do not believe REG_EAX is a real register on x86-64 regarding 
fixed registers, is
it? 
<ahuillet> I mean are you ever gonna get_fixed_var(REG_EAX) on x86-64 as 
opposed to REG_RAX ?
<ahuillet> I don't like registers not being called by their name
<ahuillet> it's both confusing and unclear as to what register is in use on 
x86-64

Eduard: I suggested aliasing REG_EAX and REG_RAX instead of using a new 
symbolic name.
Is there anything wrong with this approach?

-- 
Greetings, 
A. Huillet

------------------------------------------------------------------------------
Stay on top of everything new and different, both inside and 
around Java (TM) technology - register by April 22, and save
$200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco.
300 plus technical and hands-on sessions. Register today. 
Use priority code J9JMT32. http://p.sf.net/sfu/p
_______________________________________________
Jatovm-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/jatovm-devel

Reply via email to