[PATCH] Allow reference form for DW_AT_associated and DW_AT_allocated attributes

[Andrew Burgess]

* Sharma, Alok Kumar <AlokKumar.Sharma@amd.com> [2020-07-01 09:37:37 +0000]:

> Thanks for your comment/suggestion. Please find updated patch as an attachment.
> 
> gdb/ChangeLog
> 
>        * dwarf2/read.c (set_die_type): Removed conditions to restrict
>        forms for DW_AT_associated and DW_AT_allocated attributes,
>        which is already checked in function attr_to_dynamic_prop.
> 
> Regards,
> Alok

Looks good.  Please go ahead and apply.

Thanks,
Andrew


> 
> -----Original Message-----
> From: Andrew Burgess <andrew.burgess@embecosm.com> 
> Sent: Wednesday, July 1, 2020 1:50 PM
> To: Sharma, Alok Kumar <AlokKumar.Sharma@amd.com>
> Cc: gdb-patches@sourceware.org; George, Jini Susan <JiniSusan.George@amd.com>; Achra, Nitika <Nitika.Achra@amd.com>
> Subject: Re: [PATCH] Allow reference form for DW_AT_associated and DW_AT_allocated attributes
> 
> [CAUTION: External Email]
> 
> * Sharma, Alok Kumar <AlokKumar.Sharma@amd.com> [2020-07-01 07:01:53 +0000]:
> 
> > Hi All,
> >
> > I request you all to please review this patch. Below are the details.
> >
> > Problem Description:
> >     Currently, GDB rejects the (die) reference form while it accepts exprloc
> >     form. It is allowed in DWARF standard. "Table 7.5: Attribute encodings"
> >     in DWARF5 standard. Flang compiler assigns (die) reference to
> >     DW_AT_associated and DW_AT_allocated for some cases.
> > Resolution:
> >     Now ' set_die_type' Is modified to accept reference form as well.
> > This change will not impact testsuite for gfortran, and will fix 
> > failures for Flang compiler with existing testsuite.
> >
> >     gdb/ChangeLog
> >
> >             * dwarf2/read.c (set_die_type): Allow reference form for
> >             DW_AT_associated and DW_AT_allocated attributes.
> >
> > Please let me know your valuable comments.
> >
> > Regards,
> > Alok
> >
> > From 9a37e9973e513738234b4fed4b66d0ffc1438398 Mon Sep 17 00:00:00 2001
> > From: alosharm <AlokKumar.Sharma@amd.com>
> > Date: Wed, 1 Jul 2020 11:53:09 +0530
> > Subject: [PATCH] Allow reference form for DW_AT_associated and 
> > DW_AT_allocated  attributes
> >
> > Currently, GDB rejects the (die) reference form while it accepts 
> > exprloc form. It is allowed in DWARF standard. "Table 7.5: Attribute encodings"
> > in DWARF5 standard. Flang compiler assigns (die) reference to 
> > DW_AT_associated and DW_AT_allocated for some cases.
> >
> > gdb/ChangeLog
> >
> >       * dwarf2/read.c (set_die_type): Allow reference form for
> >       DW_AT_associated and DW_AT_allocated attributes.
> > ---
> >  gdb/ChangeLog     | 5 +++++
> >  gdb/dwarf2/read.c | 4 ++--
> >  2 files changed, 7 insertions(+), 2 deletions(-)
> >
> > diff --git a/gdb/ChangeLog b/gdb/ChangeLog index 
> > d64f572e07..75ea79471c 100644
> > --- a/gdb/ChangeLog
> > +++ b/gdb/ChangeLog
> > @@ -1,3 +1,8 @@
> > +2020-07-01  Alok Kumar Sharma  <AlokKumar.Sharma@amd.com>
> > +
> > +     * dwarf2/read.c (set_die_type): Allow reference form for
> > +     DW_AT_associated and DW_AT_allocated attributes.
> > +
> >  2020-06-30  Tom Tromey  <tromey@adacore.com>
> >
> >       * dwarf2/read.c (quirk_rust_enum): Correctly call diff --git 
> > a/gdb/dwarf2/read.c b/gdb/dwarf2/read.c index b097f624b6..0fba590184 
> > 100644
> > --- a/gdb/dwarf2/read.c
> > +++ b/gdb/dwarf2/read.c
> > @@ -23859,7 +23859,7 @@ set_die_type (struct die_info *die, struct 
> > type *type, struct dwarf2_cu *cu)
> >
> >    /* Read DW_AT_allocated and set in type.  */
> >    attr = dwarf2_attr (die, DW_AT_allocated, cu);
> > -  if (attr != NULL && attr->form_is_block ())
> > +  if (attr != NULL && (attr->form_is_block () || attr->form_is_ref 
> > + ()))
> >      {
> >        struct type *prop_type = cu->addr_sized_int_type (false);
> >        if (attr_to_dynamic_prop (attr, die, cu, &prop, prop_type))
> 
> Looking at attr_to_dynamic_prop we handle block, ref, and constants, otherwise we give complain about an unexpected attribute type.
> 
> I think then that there's no need to check the form of the attribute here at all, you could just perform a NULL check and proceed to handle the attribute.  The error in the else block (about incorrect attribute
> form) can be deleted.
> 
> This applies to both the changes in this patch.
> 
> Thanks,
> Andrew
> 
> 
> > @@ -23874,7 +23874,7 @@ set_die_type (struct die_info *die, struct 
> > type *type, struct dwarf2_cu *cu)
> >
> >    /* Read DW_AT_associated and set in type.  */
> >    attr = dwarf2_attr (die, DW_AT_associated, cu);
> > -  if (attr != NULL && attr->form_is_block ())
> > +  if (attr != NULL && (attr->form_is_block () || attr->form_is_ref 
> > + ()))
> >      {
> >        struct type *prop_type = cu->addr_sized_int_type (false);
> >        if (attr_to_dynamic_prop (attr, die, cu, &prop, prop_type))
> > --
> > 2.17.1
>