mtfsf: fix FPSCR_VX and FPSCR_FEX computation

Hi all,

The patch below fix the computation of FPSCR_VX and FPSCR_FEX when using the mtfsf instruction. As stated in the PowerPC manual the mtfsf instruction can't alter those bit, and thus it should always be computed.

Index: target-ppc/op_helper.c
===================================================================
RCS file: /sources/qemu/qemu/target-ppc/op_helper.c,v retrieving revision 1.73
diff -u -d -p -r1.73 op_helper.c

op_helper.c +4 -0

@@ -888,12 +888,16 @@ void do_store_fpscr (uint32_t mask)
     /* Update VX and FEX */
     if (fpscr_ix != 0)
         env->fpscr |= 1 << FPSCR_VX;
+    else
+        env->fpscr &= ~(1 << FPSCR_VX);
     if ((fpscr_ex & fpscr_eex) != 0) {
         env->fpscr |= 1 << FPSCR_FEX;
         env->exception_index = POWERPC_EXCP_PROGRAM;
         /* XXX: we should compute it properly */
         env->error_code = POWERPC_EXCP_FP;
     }
+    else
+        env->fpscr &= ~(1 << FPSCR_FEX);
     fpscr_set_rounding_mode();
 }
 #undef WORD0

Quoted 1 line · Aurelien Jarno

Hi all,

Hi,

Quoted 4 lines

The patch below fix the computation of FPSCR_VX and FPSCR_FEX when using the mtfsf instruction. As stated in the PowerPC manual the mtfsf instruction can't alter those bit, and thus it should always be computed.

You're right, the values are not computed properly. I'll commit your fix.