[2/5] gdb: add inferior parameter to set_target_gdbarch, rename to set_inferior_gdbarch

  From: Simon Marchi <simon.marchi@polymtl.ca>

Add an inferior parameter, so it doesn't depend on the current
inferior.  While at it, rename to set_inferior_gdbarch, I think that's a
better name than "target".

The sole observer of the architecture_changed observable,
pyuw_on_new_gdbarch, doesn't seem to depend on the current inferior.
Neither does registers_changed.

Update callers to pass the current inferior, no change in behavior is
expected.

Change-Id: I276e28eafd4740c94bc5233c81a86c01b4a6ae90
---
 gdb/arch-utils.c | 16 ++++++++--------
 gdb/gdbarch.h    |  4 ++--
 2 files changed, 10 insertions(+), 10 deletions(-)
  

On Thu, Nov 24, 2022 at 11:04:25AM -0500, Simon Marchi via Gdb-patches wrote:
> From: Simon Marchi <simon.marchi@polymtl.ca>
> 
> Add an inferior parameter, so it doesn't depend on the current
> inferior.  While at it, rename to set_inferior_gdbarch, I think that's a
> better name than "target".
> 
> The sole observer of the architecture_changed observable,
> pyuw_on_new_gdbarch, doesn't seem to depend on the current inferior.
> Neither does registers_changed.

Hi,

I kind of feel that it would make sense to also change the type of the
architecture_changed observable so the observers receive the inferior *
and the gdbarch *.  It will not be used at the moment, but makes it
clearer that this observable gets triggerd for a given inferior.

WDYT?

Best,
Lancelot.

> 
> Update callers to pass the current inferior, no change in behavior is
> expected.
> 
> Change-Id: I276e28eafd4740c94bc5233c81a86c01b4a6ae90
> ---
>  gdb/arch-utils.c | 16 ++++++++--------
>  gdb/gdbarch.h    |  4 ++--
>  2 files changed, 10 insertions(+), 10 deletions(-)
> 
> diff --git a/gdb/arch-utils.c b/gdb/arch-utils.c
> index dc67c632155..92caa5c3c4a 100644
> --- a/gdb/arch-utils.c
> +++ b/gdb/arch-utils.c
> @@ -617,7 +617,7 @@ gdbarch_update_p (struct gdbarch_info info)
>  		"New architecture %s (%s) selected\n",
>  		host_address_to_string (new_gdbarch),
>  		gdbarch_bfd_arch_info (new_gdbarch)->printable_name);
> -  set_target_gdbarch (new_gdbarch);
> +  set_inferior_gdbarch (current_inferior (), new_gdbarch);
>  
>    return 1;
>  }
> @@ -649,7 +649,7 @@ set_gdbarch_from_file (bfd *abfd)
>  
>    if (gdbarch == NULL)
>      error (_("Architecture of file not recognized."));
> -  set_target_gdbarch (gdbarch);
> +  set_inferior_gdbarch (current_inferior (), gdbarch);
>  }
>  
>  /* Initialize the current architecture.  Update the ``set
> @@ -1427,15 +1427,15 @@ gdbarch_find_by_info (struct gdbarch_info info)
>    return new_gdbarch;
>  }
>  
> -/* Make the specified architecture current.  */
> +/* See gdbarch.h.  */
>  
>  void
> -set_target_gdbarch (struct gdbarch *new_gdbarch)
> +set_inferior_gdbarch (inferior *inf, gdbarch *gdbarch)
>  {
> -  gdb_assert (new_gdbarch != NULL);
> -  gdb_assert (new_gdbarch->initialized_p);
> -  current_inferior ()->gdbarch = new_gdbarch;
> -  gdb::observers::architecture_changed.notify (new_gdbarch);
> +  gdb_assert (gdbarch != nullptr);
> +  gdb_assert (gdbarch->initialized_p);
> +  inf->gdbarch = gdbarch;
> +  gdb::observers::architecture_changed.notify (gdbarch);
>    registers_changed ();
>  }
>  
> diff --git a/gdb/gdbarch.h b/gdb/gdbarch.h
> index f2ba5f97765..76ffddfe0ff 100644
> --- a/gdb/gdbarch.h
> +++ b/gdb/gdbarch.h
> @@ -342,9 +342,9 @@ extern int gdbarch_update_p (struct gdbarch_info info);
>  extern struct gdbarch *gdbarch_find_by_info (struct gdbarch_info info);
>  
>  
> -/* Helper function.  Set the target gdbarch to "gdbarch".  */
> +/* Set INF's the target gdbarch to "gdbarch".  */
>  
> -extern void set_target_gdbarch (struct gdbarch *gdbarch);
> +extern void set_inferior_gdbarch (inferior *inf, gdbarch *gdbarch);
>  
>  
>  /* A registry adaptor for gdbarch.  This arranges to store the
> -- 
> 2.37.3
>
  

On 11/24/22 11:42, Lancelot SIX wrote:
> On Thu, Nov 24, 2022 at 11:04:25AM -0500, Simon Marchi via Gdb-patches wrote:
>> From: Simon Marchi <simon.marchi@polymtl.ca>
>>
>> Add an inferior parameter, so it doesn't depend on the current
>> inferior.  While at it, rename to set_inferior_gdbarch, I think that's a
>> better name than "target".
>>
>> The sole observer of the architecture_changed observable,
>> pyuw_on_new_gdbarch, doesn't seem to depend on the current inferior.
>> Neither does registers_changed.
> 
> Hi,
> 
> I kind of feel that it would make sense to also change the type of the
> architecture_changed observable so the observers receive the inferior *
> and the gdbarch *.  It will not be used at the moment, but makes it
> clearer that this observable gets triggerd for a given inferior.
> 
> WDYT?

That's a good idea, I'll add a patch that does that to my next version.

Simon