math/mpx-mul4-amd64-sse2.S: Improve the end-of-loop condition testing.

[catacomb] / math / mpx-mul4-amd64-sse2.S

diff --git a/math/mpx-mul4-amd64-sse2.S b/math/mpx-mul4-amd64-sse2.S
index da3e6d6..1c205a7 100644 (file)
--- a/math/mpx-mul4-amd64-sse2.S
+++ b/math/mpx-mul4-amd64-sse2.S
@@ -1270,10 +1270,10 @@ FUNC(mpxmont_redc4_amd64_sse2)
        jb      7b
 
        // All done for this iteration.  Start the next.
-8:     mov     rdi, DV                 // -> Z = dv[i]
-       mov     rbx, NV                 // -> X = nv[0]
-       cmp     rdi, DVLO               // all done yet?
+       cmp     DV, DVLO                // all done yet?
        jae     9f
+       mov     rdi, DV                 // -> Z = dv[i]
+       mov     rbx, NV                 // -> X = nv[0]
        add     DV, 16
        call    mont4
        add     rdi, 16