Re: [patch] infcall: Remove gdb_assert ($sp underflow)

[Jan Kratochvil <jan.kratochvil@redhat.com>]

On Sun, 28 Feb 2010 15:35:09 +0100, Daniel Jacobowitz wrote:
> On Sat, Feb 27, 2010 at 01:46:34AM +0100, Jan Kratochvil wrote:
> > +    gdb_test {sp == 0: call doubleit (1)}
> 
> Except what does this do?  I've never seen this syntax before, I'm
> surprised that it works.

It generated false PASS - not testing the internal error at all.
	Undefined command: "sp".  Try "help".
Thanks for catching it.


> I think the patch is otherwise OK.

Checked-in.


Thanks,
Jan


http://sourceware.org/ml/gdb-cvs/2010-02/msg00222.html

--- src/gdb/ChangeLog	2010/02/28 15:04:10	1.11406
+++ src/gdb/ChangeLog	2010/02/28 17:56:36	1.11407
@@ -1,3 +1,8 @@
+2010-02-28  Jan Kratochvil  <jan.kratochvil@redhat.com>
+
+	* infcall.c (call_function_by_hand): Remove gdb_assert on sp and old_sp.
+	New comment.
+
 2010-02-28  Corinna Vinschen  <vinschen@redhat.com>
 
 	* Makefile.in (SUBDIR_MI_OBS): Move mi-common.o from here...
--- src/gdb/infcall.c	2010/01/28 22:07:58	1.126
+++ src/gdb/infcall.c	2010/02/28 17:56:36	1.127
@@ -518,10 +518,9 @@
 	      /* Stack grows up.  */
 	      sp = gdbarch_frame_align (gdbarch, old_sp + 1);
 	  }
-	gdb_assert ((gdbarch_inner_than (gdbarch, 1, 2)
-		    && sp <= old_sp)
-		    || (gdbarch_inner_than (gdbarch, 2, 1)
-		       && sp >= old_sp));
+	/* SP may have underflown address zero here from OLD_SP.  Memory access
+	   functions will probably fail in such case but that is a target's
+	   problem.  */
       }
     else
       /* FIXME: cagney/2002-09-18: Hey, you loose!
--- src/gdb/testsuite/ChangeLog	2010/02/26 05:50:35	1.2154
+++ src/gdb/testsuite/ChangeLog	2010/02/28 17:56:37	1.2155
@@ -1,3 +1,8 @@
+2010-02-28  Jan Kratochvil  <jan.kratochvil@redhat.com>
+	    Daniel Jacobowitz  <dan@codesourcery.com>
+
+	* gdb.base/callfuncs.exp: New tests for $spval 0 and -1.  Remove return.
+
 2010-02-25  David S. Miller  <davem@davemloft.net>
 
 	* gdb.base/catch-syscall.exp: Allow to run on sparc*-*-linux and
--- src/gdb/testsuite/gdb.base/callfuncs.exp	2010/01/01 07:32:00	1.31
+++ src/gdb/testsuite/gdb.base/callfuncs.exp	2010/02/28 17:56:37	1.32
@@ -469,5 +469,17 @@
     fail "nested call dummies preserve register contents"
 }
 
-return 0
+# GDB should not crash by internal error on $sp underflow during the inferior
+# call.  It is OK it will stop on some: Cannot access memory at address 0x$hex.
 
+if {![target_info exists gdb,nosignals] && ![istarget "*-*-uclinux*"]} {
+    gdb_test {set $old_sp = $sp}
+
+    gdb_test {set $sp = 0}
+    gdb_test "call doubleit (1)" "" "sp == 0: call doubleit (1)"
+
+    gdb_test {set $sp = -1}
+    gdb_test "call doubleit (1)" "" "sp == -1: call doubleit (1)"
+
+    gdb_test {set $sp = $old_sp}
+}