[PATCH,sim] properly handle sync instruction variants

[PATCH,sim] properly handle sync instruction variants

[Nathan Froyd]

Recent Power ISAs have included the ability to control the type of
memory accesses ordered by the `sync' instruction by means of a two-bit
field.  The patch below modifies the ppc simulator's instruction
description to take note of this two-bit field.

Tested on powerpc-none-eabi, where it fixes ~140 g++ testsuite failures
and a handful of gcc failures.  OK to commit?

-Nathan

cvs diff: Diffing .
Index: ChangeLog
===================================================================
RCS file: /cvs/src/src/sim/ppc/ChangeLog,v
retrieving revision 1.87
diff -u -u -r1.87 ChangeLog
--- ChangeLog	15 Dec 2008 19:48:06 -0000	1.87
+++ ChangeLog	12 Jan 2009 18:04:27 -0000
@@ -1,3 +1,7 @@
+2009-01-12  Nathan Froyd  <froydnj@codesourcery.com>
+
+	* ppc-instructions (sync): Add L field.
+
 2008-12-15 Joel Sherrill <joel.sherrill@oarcorp.com>
 
 	* ppc-instructions, ppc-spr-table: Add ability
Index: ppc-instructions
===================================================================
RCS file: /cvs/src/src/sim/ppc/ppc-instructions,v
retrieving revision 1.10
diff -u -u -r1.10 ppc-instructions
--- ppc-instructions	15 Dec 2008 19:48:06 -0000	1.10
+++ ppc-instructions	12 Jan 2009 18:04:27 -0000
@@ -2494,7 +2494,7 @@
 	}
 	PPC_INSN_INT(0, (RA_BITMASK & ~1) | RB_BITMASK | RS_BITMASK, 1/*Rc*/);
 
-0.31,6./,11./,16./,21.598,31./:X::sync:Synchronize
+0.31,6./,9.L,11./,16./,21.598,31./:X::sync:Synchronize
 *601: PPC_UNIT_IU,    PPC_UNIT_IU,    1,  1,  0
 *603: PPC_UNIT_SRU,   PPC_UNIT_SRU,   1,  1,  0
 *603e:PPC_UNIT_SRU,   PPC_UNIT_SRU,   1,  1,  0

Re: [PATCH,sim] properly handle sync instruction variants

[Daniel Jacobowitz]

On Mon, Jan 12, 2009 at 10:09:54AM -0800, Nathan Froyd wrote:
> Recent Power ISAs have included the ability to control the type of
> memory accesses ordered by the `sync' instruction by means of a two-bit
> field.  The patch below modifies the ppc simulator's instruction
> description to take note of this two-bit field.
> 
> Tested on powerpc-none-eabi, where it fixes ~140 g++ testsuite failures
> and a handful of gcc failures.  OK to commit?

OK.

-- 
Daniel Jacobowitz
CodeSourcery