[PATCH] Add CTF support to GDB [2/4] Fix a bug of function dwarf_expr_frame_base

[PATCH] Add CTF support to GDB [2/4] Fix a bug of function dwarf_expr_frame_base

[Hui Zhu]

[-- Attachment #1: Type: text/plain, Size: 459 bytes --]

After I add prev patch to GDB, the GDB got crash when a trace frame didn't include all the variables that it should include.
I think this issue should be handle because the trace frame didn't include all the variables will be always happen (What I use is GDB and gdbserver).
So I post this patch to handle it.

Thanks,
Hui

2012-11-20  Hui Zhu  <hui_zhu@mentor.com>

	* dwarf2loc.c (dwarf_expr_frame_base): Add check for the return
	value of get_frame_block.

[-- Attachment #2: check-get_frame_block-return.txt --]
[-- Type: text/plain, Size: 801 bytes --]

--- a/dwarf2loc.c
+++ b/dwarf2loc.c
@@ -332,11 +332,15 @@ dwarf_expr_frame_base (void *baton, cons
      this_base method.  */
   struct symbol *framefunc;
   struct dwarf_expr_baton *debaton = (struct dwarf_expr_baton *) baton;
+  struct block *bl = get_frame_block (debaton->frame, NULL);
+
+  if (bl == NULL)
+    error (_("No symbol table info available.\n"));
 
   /* Use block_linkage_function, which returns a real (not inlined)
      function, instead of get_frame_function, which may return an
      inlined function.  */
-  framefunc = block_linkage_function (get_frame_block (debaton->frame, NULL));
+  framefunc = block_linkage_function (bl);
 
   /* If we found a frame-relative symbol then it was certainly within
      some function associated with a frame. If we can't find the frame,

Re: [PATCH] Add CTF support to GDB [2/4] Fix a bug of function dwarf_expr_frame_base

[Tom Tromey]

>>>>> "Hui" == Hui Zhu <hui_zhu@mentor.com> writes:

Hui> 2012-11-20  Hui Zhu  <hui_zhu@mentor.com>
Hui> 	* dwarf2loc.c (dwarf_expr_frame_base): Add check for the return
Hui> 	value of get_frame_block.

I think this is reasonable even without the rest of your patches.

One nit:

Hui> +  if (bl == NULL)
Hui> +    error (_("No symbol table info available.\n"));
 
I don't think this is a very clear error message.
I'm not sure what else to suggest though.

Also, no \n at the end of the error message.

Tom

Re: [PATCH] Add CTF support to GDB [2/4] Fix a bug of function dwarf_expr_frame_base

[Hui Zhu]

On Fri, Nov 30, 2012 at 3:14 AM, Tom Tromey <tromey@redhat.com> wrote:
>>>>>> "Hui" == Hui Zhu <hui_zhu@mentor.com> writes:
>
> Hui> 2012-11-20  Hui Zhu  <hui_zhu@mentor.com>
> Hui>    * dwarf2loc.c (dwarf_expr_frame_base): Add check for the return
> Hui>    value of get_frame_block.
>
> I think this is reasonable even without the rest of your patches.

OK.

>
> One nit:
>
> Hui> +  if (bl == NULL)
> Hui> +    error (_("No symbol table info available.\n"));
>
> I don't think this is a very clear error message.
> I'm not sure what else to suggest though.
>
> Also, no \n at the end of the error message.
>
> Tom

What about I change it to "frame address is not available." because in
get_frame_block:
  if (!get_frame_address_in_block_if_available (frame, &pc))
    return NULL;

Post a new version.

Thanks,
Hui

2012-11-01  Hui Zhu  <hui_zhu@mentor.com>

	* dwarf2loc.c (dwarf_expr_frame_base): Add check for the return
	value of get_frame_block.

--- a/dwarf2loc.c
+++ b/dwarf2loc.c
@@ -332,11 +332,15 @@ dwarf_expr_frame_base (void *baton, cons
      this_base method.  */
   struct symbol *framefunc;
   struct dwarf_expr_baton *debaton = (struct dwarf_expr_baton *) baton;
+  struct block *bl = get_frame_block (debaton->frame, NULL);
+
+  if (bl == NULL)
+    error (_("frame address is not available."));

   /* Use block_linkage_function, which returns a real (not inlined)
      function, instead of get_frame_function, which may return an
      inlined function.  */
-  framefunc = block_linkage_function (get_frame_block (debaton->frame, NULL));
+  framefunc = block_linkage_function (bl);

   /* If we found a frame-relative symbol then it was certainly within
      some function associated with a frame. If we can't find the frame,

Re: [PATCH] Add CTF support to GDB [2/4] Fix a bug of function dwarf_expr_frame_base

[Tom Tromey]

>>>>> "Hui" == Hui Zhu <teawater@gmail.com> writes:

Hui> What about I change it to "frame address is not available." because in
Hui> get_frame_block:
Hui>   if (!get_frame_address_in_block_if_available (frame, &pc))
Hui>     return NULL;

Seems reasonable to me.
This is ok.

Tom

Re: [PATCH] Add CTF support to GDB [2/4] Fix a bug of function dwarf_expr_frame_base

[Hui Zhu]

On Tue, Dec 4, 2012 at 3:11 AM, Tom Tromey <tromey@redhat.com> wrote:
>>>>>> "Hui" == Hui Zhu <teawater@gmail.com> writes:
>
> Hui> What about I change it to "frame address is not available." because in
> Hui> get_frame_block:
> Hui>   if (!get_frame_address_in_block_if_available (frame, &pc))
> Hui>     return NULL;
>
> Seems reasonable to me.
> This is ok.
>
> Tom

Should I commit this patch?

Thanks,
Hui

Re: [PATCH] Add CTF support to GDB [2/4] Fix a bug of function dwarf_expr_frame_base

[Tom Tromey]

>>>>> "Hui" == Hui Zhu <teawater@gmail.com> writes:

Hui> Should I commit this patch?

Sure, thanks.

Tom

Re: [PATCH] Add CTF support to GDB [2/4] Fix a bug of function dwarf_expr_frame_base

[Hui Zhu]

On Thu, Dec 6, 2012 at 12:26 AM, Tom Tromey <tromey@redhat.com> wrote:
>>>>>> "Hui" == Hui Zhu <teawater@gmail.com> writes:
>
> Hui> Should I commit this patch?
>
> Sure, thanks.
>
> Tom

http://sourceware.org/ml/gdb-cvs/2012-12/msg00021.html
Checked in.

Thanks,
Hui