[applied mips sim patch] trace "d" shift instructions.

[applied mips sim patch] trace "d" shift instructions.

[cgd]

When looking through my list of old patches, I noticed that in the
last few days i'd gotten around to putting in some changes which i'd
submitted back in 12/2002...  (I'd forgotten about my old patches.)

That led me to wondering what else of those i was missing.  This is
one chunk.

(verified that various sims still build w/ it.)

===================================================================
2002-03-01  Chris Demetriou  <cgd@broadcom.com>

	* mips.igen (do_dsll, do_dsllv, DSLL32, do_dsra, DSRA32, do_dsrl,
	DSRL32, do_dsrlv): Trace inputs and results.

Index: mips.igen
===================================================================
RCS file: /cvs/src/src/sim/mips/mips.igen,v
retrieving revision 1.22
diff -u -r1.22 mips.igen
--- mips.igen	2002/03/01 19:55:42	1.22
+++ mips.igen	2002/03/01 23:23:34
@@ -1180,13 +1180,17 @@
 
 :function:::void:do_dsll:int rt, int rd, int shift
 {
+  TRACE_ALU_INPUT2 (GPR[rt], shift);
   GPR[rd] = GPR[rt] << shift;
+  TRACE_ALU_RESULT (GPR[rd]);
 }
 
 :function:::void:do_dsllv:int rs, int rt, int rd
 {
   int s = MASKED64 (GPR[rs], 5, 0);
+  TRACE_ALU_INPUT2 (GPR[rt], s);
   GPR[rd] = GPR[rt] << s;
+  TRACE_ALU_RESULT (GPR[rd]);
 }
 
 
@@ -1213,7 +1217,9 @@
 {
   int s = 32 + SHIFT;
   check_u64 (SD_, instruction_0);
+  TRACE_ALU_INPUT2 (GPR[RT], s);
   GPR[RD] = GPR[RT] << s;
+  TRACE_ALU_RESULT (GPR[RD]);
 }
 
 000000,5.RS,5.RT,5.RD,00000,010100:SPECIAL:64::DSLLV
@@ -1230,7 +1236,9 @@
 
 :function:::void:do_dsra:int rt, int rd, int shift
 {
+  TRACE_ALU_INPUT2 (GPR[rt], shift);
   GPR[rd] = ((signed64) GPR[rt]) >> shift;
+  TRACE_ALU_RESULT (GPR[rd]);
 }
 
 
@@ -1257,7 +1265,9 @@
 {
   int s = 32 + SHIFT;
   check_u64 (SD_, instruction_0);
+  TRACE_ALU_INPUT2 (GPR[RT], s);
   GPR[RD] = ((signed64) GPR[RT]) >> s;
+  TRACE_ALU_RESULT (GPR[RD]);
 }
 
 
@@ -1283,7 +1293,9 @@
 
 :function:::void:do_dsrl:int rt, int rd, int shift
 {
+  TRACE_ALU_INPUT2 (GPR[rt], shift);
   GPR[rd] = (unsigned64) GPR[rt] >> shift;
+  TRACE_ALU_RESULT (GPR[rd]);
 }
 
 
@@ -1310,14 +1322,18 @@
 {
   int s = 32 + SHIFT;
   check_u64 (SD_, instruction_0);
+  TRACE_ALU_INPUT2 (GPR[RT], s);
   GPR[RD] = (unsigned64) GPR[RT] >> s;
+  TRACE_ALU_RESULT (GPR[RD]);
 }
 
 
 :function:::void:do_dsrlv:int rs, int rt, int rd
 {
   int s = MASKED64 (GPR[rs], 5, 0);
+  TRACE_ALU_INPUT2 (GPR[rt], s);
   GPR[rd] = (unsigned64) GPR[rt] >> s;
+  TRACE_ALU_RESULT (GPR[rd]);
 }
 
 

Re: [applied mips sim patch] trace "d" shift instructions.

[cgd]

Or, it will be applied shortly.  "technical difficulties."



chris

Re: [applied mips sim patch] trace "d" shift instructions.

[Eli Zaretskii]

> From: cgd@broadcom.com
> Date: 01 Mar 2002 15:26:17 -0800
> 
> ===================================================================
> 2002-03-01  Chris Demetriou  <cgd@broadcom.com>
> 
> 	* mips.igen (do_dsll, do_dsllv, DSLL32, do_dsra, DSRA32, do_dsrl,
> 	DSRL32, do_dsrlv): Trace inputs and results.

Just a quick note: this is not how ChangeLog entries that span several
lines should be formatted, according to standards.texi.  You should do
it like this:

	* mips.igen (do_dsll, do_dsllv, DSLL32, do_dsra, DSRA32, do_dsrl)
	(DSRL32, do_dsrlv): Trace inputs and results.

That is, close the parens on each line and open them anew on the next
one.  One reason for this format is that Emacs highlights parts of the
entries in different colors, and it depends on this format to DTRT.
Another reason is that it makes it easier to grep ChangeLog files for
function/variable names.

If you use Emacs 21.1 or later, the command "C-x 4 a", when typed
inside the function/variable definition for which you need to write a
ChangeLog entry, will do the above formatting automatically for you.

Re: [applied mips sim patch] trace "d" shift instructions.

[cgd]

At Sun, 3 Mar 2002 13:52:35 +0200 (IST), Eli Zaretskii wrote:
> Just a quick note: this is not how ChangeLog entries that span several
> lines should be formatted, according to standards.texi.  You should do
> it like this:
> [ ... ]

Hmm, I don't think anybody's mentioned that to me previously.  Thanks
for pointing it out.



chris