This is the mail archive of the gdb-patches@sourceware.cygnus.com mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: Unifying i387 support



> >> Does anyone care for the MMX view of the ST(i) registers?  Perhaps, if
> >> we design a change in the layout, it would be better to make it
> >> last...
> 
> Jim> I don't understand.  They do appear last.
> 
> I think Eli was asking whether it would be better to use the FXSAVE
> format introduced with MMX and extended with the SSE instead of the
> FSAVE format described above.  It might be worth doing if we're 
> going to support MMX and SSE registers.  

Oh, Eli meant, "make it a good long-term change"!!!  *blush*

My Pentium manual predates MMX, so I can't really comment much on
this.  It looks like the FXSAVE instruction saves the floating-point
registers separately from the XMM registers; doesn't that mean it's
saving the same bits twice?

As I understand it, there are multiple ways to interpret the contents
of an MMX register.  It might contain one double, or two floats.  And
only the user knows what is appropriate at any given time.  I can't
tell if this impinges on our choice of register layout.  Does anyone
have any idea how to handle this?

I guess I would assign distinct register numbers and names to each
possible component --- ie. $mmx4d, a double, might overlap with
$mmx4f0 and $mmx4f1.  The virtual type of $mmx4d would be double,
while $mmx4f0 and $mmx4f1 would be floats.  But I don't know if GDB
handles overlapping registers.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]