Re: [PATCH 2/2 ver 3] PowerPC, fix support for printing the function return value for non-trivial values.

[Carl Love <cel@us.ibm.com>]

Tom, Ulrich:

On Tue, 2022-11-15 at 10:16 +0000, Ulrich Weigand wrote:
> Tom de Vries <tdevries@suse.de> wrote:
> > On 11/14/22 17:47, Ulrich Weigand via Gdb-patches wrote:
> > > I believe all of Kevin's and Bruno's comments have been addressed
> > > in this version.  I've reviewed it myself as well, and it looks
> > > good to me.
> > > This is OK.
> > On x86_64-linux, I run into a segfault:
> > [...]
> > (gdb) up
> > #1  0x0000000000852d94 in finish_command (arg=0x0, from_tty=0)
> >     at /home/vries/gdb_versions/devel/src/gdb/infcmd.c:1887
> > 1887        = check_typedef (sm->function->type ()->target_type
> > ());
> > (gdb) p sm->function
> > $1 = (symbol *) 0x0
> 
> Ah, I missed that possibility, sorry.
> 
> Carl, if sm->function is NULL, then the whole check needs
> to be skipped, and sm->return_buf should be set to 0.
> 
> Can you come up with a fix along those lines (or else,
> just revert the patch for now)?   Thanks!
> 

I tested the last mainline code with the command on an X86-64 box.

   make check RUNTESTFLAGS='GDB=/home/carll/bin/gdb  gdb.asm/asm-source.exp  '

Which generated an error.  I then ran the command, from Tom's email:

 gdb -q -batch -x outputs/gdb.asm/asm-source/gdb.in.1

from directory gdb/testsuite which generated a segmentation fault as
expected.

Once I applied the following patch, recompiled and retested, the make
check seemed to run fine and the gdb -q batch test ran without
generating a segmentation fault.

The patch appears to fix the issue.  Not sure why this didn't show up
in my original X86 testing?  I will go back and look some more to see
if I can track down why I didn't see the issue.

Sorry for the bug.  

Tom, can you take a look at the patch and try it out to make sure it
fixes the issues on your system.  Thanks.

                  Carl Love

----------------------------------------------------------------
Bug fix in commit for printing the function return value for non-trivial values

The recent commit:

  commit a0eda3df5b750ae32576a9be092b361281a41787
  Author: Carl Love <cel@us.ibm.com>
  Date:   Mon Nov 14 16:22:37 2022 -0500

    PowerPC, fix support for printing the function return value for non-trivial values.

Is generating a segmentation fault on x86_64-linux.

  segfault:
  ...
  PASS: gdb.asm/asm-source.exp: info source asmsrc1.s
  ERROR: GDB process no longer exists
  UNRESOLVED: gdb.asm/asm-source.exp: finish from foo3
  ...

  Reproduced on command line:
  ...
  $ gdb -q -batch -x outputs/gdb.asm/asm-source/gdb.in.1
  ...

  The problem seems to be that:
  ...
  Thread 1 "gdb" received signal SIGSEGV, Segmentation fault.
  0x000000000043de7a in symbol::type (this=0x0) at
  .../gdb_versions/devel/src/gdb/symtab.h:1287
  1287        return m_type;
  ...
  because:
  ...
  (gdb) up
  #1  0x0000000000852d94 in finish_command (arg=0x0, from_tty=0)
     at .../gdb_versions/devel/src/gdb/infcmd.c:1887
  1887        = check_typedef (sm->function->type ()->target_type ());
  (gdb) p sm->function
  $1 = (symbol *) 0x0

The code is not checking if sm->function is NULL.  If sm->function is NULL
the check for the return buffer should be skipped.
---
 gdb/infcmd.c | 31 ++++++++++++++++++-------------
 1 file changed, 18 insertions(+), 13 deletions(-)

diff --git a/gdb/infcmd.c b/gdb/infcmd.c
index b71dc10370b..2876f6d9c9d 100644
--- a/gdb/infcmd.c
+++ b/gdb/infcmd.c
@@ -1880,23 +1880,28 @@ finish_command (const char *arg, int from_tty)
 
   /* Determine the return convention.  If it is RETURN_VALUE_STRUCT_CONVENTION,
      attempt to determine the address of the return buffer.  */
-  enum return_value_convention return_value;
-  struct gdbarch *gdbarch = get_frame_arch (callee_frame);
+  if (sm->function != NULL)
+    {
+      enum return_value_convention return_value;
+      struct gdbarch *gdbarch = get_frame_arch (callee_frame);
 
-  struct type * val_type
-    = check_typedef (sm->function->type ()->target_type ());
+      struct type * val_type
+	= check_typedef (sm->function->type ()->target_type ());
 
-  return_value = gdbarch_return_value (gdbarch,
-				       read_var_value (sm->function, NULL,
-						       callee_frame),
-				       val_type, NULL, NULL, NULL);
+      return_value = gdbarch_return_value (gdbarch,
+					   read_var_value (sm->function, NULL,
+							   callee_frame),
+					   val_type, NULL, NULL, NULL);
 
-  if (return_value == RETURN_VALUE_STRUCT_CONVENTION
-      && val_type->code () != TYPE_CODE_VOID)
-    sm->return_buf = gdbarch_get_return_buf_addr (gdbarch, val_type,
-						  callee_frame);
+      if (return_value == RETURN_VALUE_STRUCT_CONVENTION
+	  && val_type->code () != TYPE_CODE_VOID)
+	sm->return_buf = gdbarch_get_return_buf_addr (gdbarch, val_type,
+						      callee_frame);
+      else
+	sm->return_buf = 0;
+    }
   else
-    sm->return_buf = 0;
+    sm->return_buf = 0;  /* Return buffer address is not available.  */
 
   /* Print info on the selected frame, including level number but not
      source.  */
-- 
2.31.1