From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 61320 invoked by alias); 7 Mar 2019 15:53:09 -0000 Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org Received: (qmail 61303 invoked by uid 89); 7 Mar 2019 15:53:08 -0000 Authentication-Results: sourceware.org; auth=none X-Spam-SWARE-Status: No, score=-25.4 required=5.0 tests=BAYES_00,GIT_PATCH_0,GIT_PATCH_1,GIT_PATCH_2,GIT_PATCH_3,KAM_STOCKGEN,SPF_HELO_PASS,SPF_PASS autolearn=ham version=3.3.1 spammy=H*r:172.16.0, Thus X-HELO: simark.ca Received: from simark.ca (HELO simark.ca) (158.69.221.121) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with ESMTP; Thu, 07 Mar 2019 15:53:06 +0000 Received: from [172.16.0.89] (192-222-157-41.qc.cable.ebox.net [192.222.157.41]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by simark.ca (Postfix) with ESMTPSA id C2C761E4A5; Thu, 7 Mar 2019 10:53:04 -0500 (EST) Subject: Re: [PATCH v2 06/11] Add a more general version of lookup_struct_elt_type. To: John Baldwin , gdb-patches@sourceware.org References: <9a5a86e3591c8fe6c0fc8efb6151547902a63d3c.1549672588.git.jhb@FreeBSD.org> From: Simon Marchi Message-ID: Date: Thu, 07 Mar 2019 15:53:00 -0000 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101 Thunderbird/60.5.1 MIME-Version: 1.0 In-Reply-To: <9a5a86e3591c8fe6c0fc8efb6151547902a63d3c.1549672588.git.jhb@FreeBSD.org> Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: 7bit X-SW-Source: 2019-03/txt/msg00149.txt.bz2 On 2019-02-08 7:40 p.m., John Baldwin wrote: > lookup_struct_elt is a new function which returns a tuple of > information about a component of a structure or union. The returned > tuple contains a pointer to the struct field object for the component > as well as a bit offset of that field within the structure. If the > field names a field in an anonymous substructure, the offset is the > "global" offset relative to the original structure type. Can you add this bit of detail (about anonymous struct) to the function doc? > If noerr is > set, then the returned tuple will set the field pointer to NULL to > indicate a missing component rather than throwing an error. > > lookup_struct_elt_type is now reimplemented in terms of this new > function. It simply returns the type of the returned field. > > gdb/ChangeLog: > > * gdbtypes.c (lookup_struct_elt): New function. > (lookup_struct_elt_type): Reimplement via lookup_struct_elt. > * gdbtypes.h (struct struct_elt): New type. > (lookup_struct_elt): New prototype. > --- > gdb/ChangeLog | 7 ++++++ > gdb/gdbtypes.c | 60 ++++++++++++++++++++++++++++++++------------------ > gdb/gdbtypes.h | 19 ++++++++++++++++ > 3 files changed, 65 insertions(+), 21 deletions(-) > > diff --git a/gdb/ChangeLog b/gdb/ChangeLog > index cebd63bcb5..c7fee7eb11 100644 > --- a/gdb/ChangeLog > +++ b/gdb/ChangeLog > @@ -1,3 +1,10 @@ > +2019-02-08 John Baldwin > + > + * gdbtypes.c (lookup_struct_elt): New function. > + (lookup_struct_elt_type): Reimplement via lookup_struct_elt. > + * gdbtypes.h (struct struct_elt): New type. > + (lookup_struct_elt): New prototype. > + > 2019-02-08 John Baldwin > > * gdbtypes.c (lookup_struct_elt_type): Update comment and > diff --git a/gdb/gdbtypes.c b/gdb/gdbtypes.c > index e4acb0e985..0f3a450f9f 100644 > --- a/gdb/gdbtypes.c > +++ b/gdb/gdbtypes.c > @@ -1644,7 +1644,8 @@ lookup_template_type (char *name, struct type *type, > return (SYMBOL_TYPE (sym)); > } > > -/* Given a type TYPE, lookup the type of the component named NAME. > +/* Given a type TYPE, lookup the field and offset of the component named > + NAME. > > TYPE can be either a struct or union, or a pointer or reference to > a struct or union. If it is a pointer or reference, its target > @@ -1652,11 +1653,11 @@ lookup_template_type (char *name, struct type *type, > as specified for the definitions of the expression element types > STRUCTOP_STRUCT and STRUCTOP_PTR. > > - If NOERR is nonzero, return NULL if there is no component named > - NAME. */ > + If NOERR is nonzero, the returned structure will have field set to > + NULL if there is no component named NAME. */ For both functions, please move the comment to the .h, and add /* See gdbtypes.h. */ here. > > -struct type * > -lookup_struct_elt_type (struct type *type, const char *name, int noerr) > +struct_elt > +lookup_struct_elt (struct type *type, const char *name, int noerr) > { > int i; > > @@ -1683,39 +1684,56 @@ lookup_struct_elt_type (struct type *type, const char *name, int noerr) > > if (t_field_name && (strcmp_iw (t_field_name, name) == 0)) > { > - return TYPE_FIELD_TYPE (type, i); > + return struct_elt (&TYPE_FIELD(type, i), TYPE_FIELD_BITPOS (type, i)); > } > else if (!t_field_name || *t_field_name == '\0') > { > - struct type *subtype > - = lookup_struct_elt_type (TYPE_FIELD_TYPE (type, i), name, 1); > - > - if (subtype != NULL) > - return subtype; > + struct_elt elt > + = lookup_struct_elt (TYPE_FIELD_TYPE (type, i), name, 1); > + if (elt.field != NULL) > + { > + elt.offset += TYPE_FIELD_BITPOS (type, i); > + return elt; > + } > } > } > > /* OK, it's not in this class. Recursively check the baseclasses. */ > for (i = TYPE_N_BASECLASSES (type) - 1; i >= 0; i--) > { > - struct type *t; > - > - t = lookup_struct_elt_type (TYPE_BASECLASS (type, i), name, 1); > - if (t != NULL) > - { > - return t; > - } > + struct_elt elt = lookup_struct_elt (TYPE_BASECLASS (type, i), name, 1); > + if (elt.field != NULL) > + return elt; > } > > if (noerr) > - { > - return NULL; > - } > + return struct_elt (); > > std::string type_name = type_to_string (type); > error (_("Type %s has no component named %s."), type_name.c_str (), name); > } > > +/* Given a type TYPE, lookup the type of the component named NAME. > + > + TYPE can be either a struct or union, or a pointer or reference to > + a struct or union. If it is a pointer or reference, its target > + type is automatically used. Thus '.' and '->' are interchangable, > + as specified for the definitions of the expression element types > + STRUCTOP_STRUCT and STRUCTOP_PTR. > + > + If NOERR is nonzero, return NULL if there is no component named > + NAME. */ > + > +struct type * > +lookup_struct_elt_type (struct type *type, const char *name, int noerr) > +{ > + struct_elt elt = lookup_struct_elt (type, name, noerr); > + if (elt.field != NULL) > + return FIELD_TYPE (*elt.field); > + else > + return NULL; > +} > + > /* Store in *MAX the largest number representable by unsigned integer type > TYPE. */ > > diff --git a/gdb/gdbtypes.h b/gdb/gdbtypes.h > index a6d4f64e9b..894c7b2fc6 100644 > --- a/gdb/gdbtypes.h > +++ b/gdb/gdbtypes.h > @@ -1873,6 +1873,25 @@ extern struct type *allocate_stub_method (struct type *); > > extern const char *type_name_or_error (struct type *type); > > +struct struct_elt > +{ > + /* The field of the element, or NULL if no element was found. */ > + struct field *field; > + > + /* The bit offset of the element in the parent structure. */ > + LONGEST offset; > + > + struct_elt () > + : field (nullptr), offset (0) > + {} > + > + struct_elt (struct field *field, LONGEST offset) > + : field (field), offset (offset) > + {} > +}; Not really a big deal, but I find it a bit overkill to define a type just for this, I would have probably just returned the the offset as an output parameter. But maybe this way is better, in that if we want to add something to the return type, we don't have to update the callers. So the patch LGTM with the nits noted above. Simon