From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 70283 invoked by alias); 31 May 2016 14:13:26 -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 70271 invoked by uid 89); 31 May 2016 14:13:25 -0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=-3.3 required=5.0 tests=BAYES_00,RP_MATCHES_RCVD,SPF_HELO_PASS autolearn=ham version=3.3.2 spammy=H*MI:sk:2016053 X-HELO: mx1.redhat.com Received: from mx1.redhat.com (HELO mx1.redhat.com) (209.132.183.28) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with (AES256-GCM-SHA384 encrypted) ESMTPS; Tue, 31 May 2016 14:13:24 +0000 Received: from int-mx09.intmail.prod.int.phx2.redhat.com (int-mx09.intmail.prod.int.phx2.redhat.com [10.5.11.22]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id 53433C0467F1; Tue, 31 May 2016 14:13:23 +0000 (UTC) Received: from [127.0.0.1] (ovpn01.gateway.prod.ext.ams2.redhat.com [10.39.146.11]) by int-mx09.intmail.prod.int.phx2.redhat.com (8.14.4/8.14.4) with ESMTP id u4VEDL6o022944; Tue, 31 May 2016 10:13:22 -0400 Subject: Re: [PATCH v3][PR gdb/19893] Fix handling of synthetic C++ references To: Martin Galvan , gdb-patches@sourceware.org References: <20160530194057.13511-1-martin.galvan@tallertechnologies.com> From: Pedro Alves Message-ID: Date: Tue, 31 May 2016 14:13:00 -0000 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Thunderbird/45.1.0 MIME-Version: 1.0 In-Reply-To: <20160530194057.13511-1-martin.galvan@tallertechnologies.com> Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: 7bit X-SW-Source: 2016-05/txt/msg00553.txt.bz2 On 05/30/2016 08:40 PM, Martin Galvan wrote: > > if (!value_bytes_available (val, embedded_offset, TYPE_LENGTH (type))) > @@ -484,6 +498,42 @@ generic_val_print_memberptr (struct type *type, const gdb_byte *valaddr, > original_value, options, 0, stream); > } > > +/* Print '@' followed by the address contained in ADDRESS_BUFFER. */ > + > +static void > +print_ref_address(struct type *type, const gdb_byte *address_buffer, > + int embedded_offset, struct ui_file *stream) Space before parenthesis. > +{ > + struct gdbarch *gdbarch = get_type_arch (type); > + > + if (address_buffer != NULL) > + { > + CORE_ADDR address > + = extract_typed_address (address_buffer + embedded_offset, type); > + > + fprintf_filtered (stream, "@"); > + fputs_filtered (paddress (gdbarch, address), stream); > + } > + /* Else: we have a non-addressable value, such as a DW_AT_const_value. */ > +} > + > +/* Return the address of the value represented by DEREF_VAL, or NULL if it's > + non-addressable. */ > + > +static const gdb_byte * > +get_address_from_deref_val (struct value *deref_val) This is returning the value contents of the pointer value, not the value's address. I think it's a confusing interface. How about: /* If VAL is addressable, return the value contents buffer of a value that represents a pointer to VAL. Otherwise return NULL. */ static const gdb_byte * get_value_addr_contents (struct value *val) > static void > @@ -492,41 +542,58 @@ generic_val_print_ref (struct type *type, const gdb_byte *valaddr, > const struct value *original_value, > const struct value_print_options *options) > { > - struct gdbarch *gdbarch = get_type_arch (type); > struct type *elttype = check_typedef (TYPE_TARGET_TYPE (type)); > + struct value *deref_val = NULL; > + const gdb_byte *address; > + const int value_is_synthetic > + = value_bits_synthetic_pointer (original_value, > + TARGET_CHAR_BIT * embedded_offset, > + TARGET_CHAR_BIT * TYPE_LENGTH (type)); > + const int must_coerce_ref = (options->addressprint && value_is_synthetic) > + || options->deref_ref; If the expression is split between lines, you need to wrap it in an extra set of parens, and reindent: const int must_coerce_ref = ((options->addressprint && value_is_synthetic) || options->deref_ref); > + const int type_is_defined = TYPE_CODE (elttype) != TYPE_CODE_UNDEF; > + > + if (must_coerce_ref && type_is_defined) > + { > + deref_val = coerce_ref_if_computed (original_value); > + > + if (deref_val != NULL) > + { > + /* More complicated computed references are not supported. */ > + gdb_assert (embedded_offset == 0); > + } > + else > + deref_val = value_at (TYPE_TARGET_TYPE (type), > + unpack_pointer (type, valaddr + embedded_offset)); > + } > + /* Else, original_value isn't a synthetic reference or we don't have to print > + the reference's contents. > + > + Notice that for references to TYPE_CODE_STRUCT, 'set print object on' will > + cause original_value to be a not_lval instead of an lval_computed, > + which will make value_bits_synthetic_pointer return false. > + This happens because if options->objectprint is true, c_value_print will > + overwrite original_value's contents with the result of coercing > + the reference through value_addr, and then set its type back to > + TYPE_CODE_REF. In that case we don't have to coerce the reference again; > + we can simply treat it as non-synthetic and move on. */ > > if (options->addressprint) > { > - CORE_ADDR addr > - = extract_typed_address (valaddr + embedded_offset, type); > + address = value_is_synthetic && type_is_defined ? > + get_address_from_deref_val (deref_val) : valaddr; Operators go on next line, not at end of line, and wrap in parens: address = (value_is_synthetic && type_is_defined ? get_address_from_deref_val (deref_val) : valaddr); Otherwise OK. Please push. Thanks, Pedro Alves