[PATCH] powerpc-prologue.exp

[PATCH] powerpc-prologue.exp

[Mark Kettenis]

While running the testsuite on openbsd4.4-unknown-powerpc, I spotted
the following new fail:

+FAIL: gdb.arch/powerpc-prologue.exp: saved registers in PIC

The problem is that "info frame" no longer reports r31 as saved.  As
far as I can tell, the assembly code doesn't actually save r31, so I
suspect this is actually a bug in the testcase that got uncovered by
the following fix by Daniel:

2008-08-18  Daniel Jacobowitz  <dan@codesourcery.com>

	* rs6000-tdep.c (struct rs6000_framedata): Add gpr_mask, used_bl,
	lr_register.
	(rs6000_in_function_epilogue_p): Check for bctr.
	(skip_prologue): Initialize lr_register.  Set lr_reg to a register
	number.  Set gpr_mask and used_bl.  Continue scanning while some
	expected registers are not saved.  Set lr_register if LR is not
	stored.
	(rs6000_frame_cache): Handle gpr_mask and lr_register.

So the following testsuite fix seems to be appropriate.  ok?


Index: ChangeLog
from  Mark Kettenis  <kettenis@gnu.org>

	* gdb.arch/powerpc-prologue.exp: Code doesn't save %r31 so don't
	expect to find it in the "info frame" output.

Index: gdb.arch/powerpc-prologue.exp
===================================================================
RCS file: /cvs/src/src/gdb/testsuite/gdb.arch/powerpc-prologue.exp,v
retrieving revision 1.6
diff -u -p -r1.6 powerpc-prologue.exp
--- gdb.arch/powerpc-prologue.exp 18 Aug 2008 12:32:47 -0000 1.6
+++ gdb.arch/powerpc-prologue.exp 20 Aug 2008 14:14:44 -0000
@@ -64,7 +64,7 @@ gdb_test "backtrace 10" \
 	"backtrace in PIC"
 
 gdb_test "info frame" \
-	".*Saved registers:.*r30 at.*r31 at.*pc at.*lr at.*" \
+	".*Saved registers:.*r30 at.*pc at.*lr at.*" \
 	"saved registers in PIC"
 
 # Testcase for scheduled prologue.

Re: [PATCH] powerpc-prologue.exp

[Daniel Jacobowitz]

On Wed, Aug 20, 2008 at 04:20:11PM +0200, Mark Kettenis wrote:
> While running the testsuite on openbsd4.4-unknown-powerpc, I spotted
> the following new fail:
> 
> +FAIL: gdb.arch/powerpc-prologue.exp: saved registers in PIC
> 
> The problem is that "info frame" no longer reports r31 as saved.  As
> far as I can tell, the assembly code doesn't actually save r31, so I
> suspect this is actually a bug in the testcase that got uncovered by
> the following fix by Daniel:

Is your testsuite directory up to date?  I think I committed this
testsuite fix at the same time as the patch - it's revision 1.6 of
that file.

-- 
Daniel Jacobowitz
CodeSourcery

Re: [PATCH] powerpc-prologue.exp

[Mark Kettenis]

> Date: Wed, 20 Aug 2008 10:45:43 -0400
> From: Daniel Jacobowitz <drow@false.org>
> 
> On Wed, Aug 20, 2008 at 04:20:11PM +0200, Mark Kettenis wrote:
> > While running the testsuite on openbsd4.4-unknown-powerpc, I spotted
> > the following new fail:
> > 
> > +FAIL: gdb.arch/powerpc-prologue.exp: saved registers in PIC
> > 
> > The problem is that "info frame" no longer reports r31 as saved.  As
> > far as I can tell, the assembly code doesn't actually save r31, so I
> > suspect this is actually a bug in the testcase that got uncovered by
> > the following fix by Daniel:
> 
> Is your testsuite directory up to date?  I think I committed this
> testsuite fix at the same time as the patch - it's revision 1.6 of
> that file.

It is.  There are two "info frame" tests though and it looks as if you
only the second one.

Re: [PATCH] powerpc-prologue.exp

[Daniel Jacobowitz]

On Wed, Aug 20, 2008 at 05:02:52PM +0200, Mark Kettenis wrote:
> It is.  There are two "info frame" tests though and it looks as if you
> only the second one.

Drat, it was added after the original posting of the patch and neither
Joel nor I noticed :-(

Thanks for fixing it!  Your patch looks correct to me.

-- 
Daniel Jacobowitz
CodeSourcery