[PATCH v2 28/42] Remove dwarf2_per_cu_data::objfile ()

[Simon Marchi]

On 2020-05-28 12:52 p.m., Simon Marchi wrote:
> On 2020-05-28 12:24 p.m., Christian Biesinger wrote:
>> On Thu, May 28, 2020 at 10:33 AM Simon Marchi <simark@simark.ca> wrote:
>>>
>>> On 2020-05-28 9:05 a.m., Tom de Vries wrote:
>>>> Hi,
>>>>
>>>> thanks for looking into this.
>>>>
>>>> I tried out the patch, and it fixes all regressions for me.
>>>>
>>>> Thanks,
>>>> - Tom
>>>
>>> Thanks, I pushed it with the following commit message.
>>
>> Maybe add a comment describing this case, so that future developers
>> looking at the code know why the two objfiles can be different?
> 
> I am the first one to ask others to comment their code and not assume that everybody
> is a domain expert, so yes :)
> 
> I would add this:
> 
> From 54c8c8f45b2196538a6c87c4cc7d498a8a5ade36 Mon Sep 17 00:00:00 2001
> From: Simon Marchi <simon.marchi@polymtl.ca>
> Date: Thu, 28 May 2020 12:38:03 -0400
> Subject: [PATCH] gdb: add comment in
>  dwarf_evaluate_loc_desc::push_dwarf_reg_entry_value
> 
> Add a comment to clarify why we temporarily override some of the
> context's fields, and especially the per_objfile field.  A longer
> explanation can be found in this previous commit
> 
>     44486dcf19b ("gdb: use caller objfile in dwarf_evaluate_loc_desc::push_dwarf_reg_entry_value")
> 
> gdb/ChangeLog:
> 
> 	* dwarf2/loc.c (class dwarf_evaluate_loc_desc)
> 	<push_dwarf_reg_entry_value>: Add comment.
> 
> Change-Id: I60c6e1062799f729b30a9db78bcb6448783324b4
> ---
>  gdb/dwarf2/loc.c | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/gdb/dwarf2/loc.c b/gdb/dwarf2/loc.c
> index 1aab1a4f51bc..400bb4d16fc7 100644
> --- a/gdb/dwarf2/loc.c
> +++ b/gdb/dwarf2/loc.c
> @@ -731,6 +731,12 @@ class dwarf_evaluate_loc_desc : public dwarf_expr_context
>        throw_error (NO_ENTRY_VALUE_ERROR,
>  		   _("Cannot resolve DW_AT_call_data_value"));
> 
> +    /* We are about to evaluate an expression in the context of the caller
> +       of the current frame.  This evaluation context may be different from
> +       the current (callee's) context), so temporarily set the caller's context.
> +
> +       It is possible for the caller to be from a different objfile from the
> +       callee if the call is made through a function pointer.  */
>      scoped_restore save_frame = make_scoped_restore (&this->frame,
>  						     caller_frame);
>      scoped_restore save_per_cu = make_scoped_restore (&this->per_cu,
> -- 
> 2.26.2
> 

I pushed this.

Simon