]> Git Repo - qemu.git/commit
target-ppc: Little Endian Correction to Load/Store Vector Element
authorTom Musta <[email protected]>
Wed, 25 Sep 2013 07:42:46 +0000 (17:42 +1000)
committerAlexander Graf <[email protected]>
Fri, 25 Oct 2013 21:25:45 +0000 (23:25 +0200)
commitbbfb6f132abc032229f5c1f25e6e959861c6f759
treeb7a4fdb8428be4ae74620c06ac7ebd88f1ac0360
parent04f1f7842e18c4b5e50203cc5b207cafb7c62974
target-ppc: Little Endian Correction to Load/Store Vector Element

The Load Vector Element (lve*x) and Store Vector Element (stve*x)
instructions not only byte-swap in Little Endian mode, they also
invert the element that is accessed. For example, the RTL for
lvehx contains this:

     eb <-- EA[60:63]
     if Big-Endian byte ordering then
         VRT[8*eb:8*eb+15] <-- MEM(EA,2)
     else
         VRT[112-(8*eb):127-(8*eb)] <-- MEM(EA,2)

This patch adds the element inversion, as described in the last line
of the RTL.

Signed-off-by: Tom Musta <[email protected]>
Reviewed-by: Anton Blanchard <[email protected]>
Signed-off-by: Alexander Graf <[email protected]>
target-ppc/mem_helper.c
This page took 0.021172 seconds and 4 git commands to generate.