From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 28007 invoked by alias); 14 Apr 2017 15:07:33 -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 27917 invoked by uid 89); 14 Apr 2017 15:07:31 -0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=-23.9 required=5.0 tests=AWL,BAYES_00,GIT_PATCH_0,GIT_PATCH_1,GIT_PATCH_2,GIT_PATCH_3,SPF_HELO_PASS,SPF_SOFTFAIL autolearn=ham version=3.3.2 spammy=2046 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; Fri, 14 Apr 2017 15:07:29 +0000 Received: by simark.ca (Postfix, from userid 33) id 756AA1E4A3; Fri, 14 Apr 2017 11:07:29 -0400 (EDT) To: Andreas Arnez Subject: Re: [PATCH 8/9] Respect piece offset for DW_OP_bit_piece X-PHP-Originating-Script: 33:rcube.php MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII; format=flowed Content-Transfer-Encoding: 7bit Date: Fri, 14 Apr 2017 15:07:00 -0000 From: Simon Marchi Cc: gdb-patches@sourceware.org In-Reply-To: <1491586736-21296-9-git-send-email-arnez@linux.vnet.ibm.com> References: <1491586736-21296-1-git-send-email-arnez@linux.vnet.ibm.com> <1491586736-21296-9-git-send-email-arnez@linux.vnet.ibm.com> Message-ID: <50c411f610271ff5c5a07b4ac10e8cb6@polymtl.ca> X-Sender: simon.marchi@polymtl.ca User-Agent: Roundcube Webmail/1.2.4 X-IsSubscribed: yes X-SW-Source: 2017-04/txt/msg00463.txt.bz2 On 2017-04-07 13:38, Andreas Arnez wrote: > So far GDB ignores the piece offset of all kinds of DWARF bit > pieces (DW_OP_bit_piece) and treats such pieces as if the offset was > zero. > > This is fixed, and an appropriate test is added. > > gdb/ChangeLog: > > * dwarf2loc.c (read_pieced_value): Respect the piece offset, as > given by DW_OP_bit_piece. > (write_pieced_value): Likewise. > > Andreas Arnez > > * gdb.dwarf2/var-access.exp: Add test for composite location with > nonzero piece offsets. > --- > gdb/dwarf2loc.c | 22 ++++++++++++++------ > gdb/testsuite/gdb.dwarf2/var-access.exp | 37 > +++++++++++++++++++++++++++++++++ > 2 files changed, 53 insertions(+), 6 deletions(-) > > diff --git a/gdb/dwarf2loc.c b/gdb/dwarf2loc.c > index 67df598..045c2d3 100644 > --- a/gdb/dwarf2loc.c > +++ b/gdb/dwarf2loc.c > @@ -1824,11 +1824,13 @@ read_pieced_value (struct value *v) > int optim, unavail; > > if (gdbarch_byte_order (arch) == BFD_ENDIAN_BIG > - && p->size < reg_bits) > + && p->offset + p->size < reg_bits) > { > /* Big-endian, and we want less than full size. */ > - source_offset_bits += reg_bits - p->size; > + source_offset_bits += reg_bits - (p->offset + p->size); > } > + else > + source_offset_bits += p->offset; > this_size = bits_to_bytes (source_offset_bits, this_size_bits); > buffer.reserve (this_size); > > @@ -1850,6 +1852,7 @@ read_pieced_value (struct value *v) > break; > > case DWARF_VALUE_MEMORY: > + source_offset_bits += p->offset; > this_size = bits_to_bytes (source_offset_bits, this_size_bits); > buffer.reserve (this_size); > > @@ -1869,12 +1872,14 @@ read_pieced_value (struct value *v) > ULONGEST obj_size = 8 * TYPE_LENGTH (value_type (p->v.value)); > > /* Use zeroes if piece reaches beyond stack value. */ > - if (p->size > obj_size) > + if (p->offset + p->size > obj_size) > break; > > /* Piece is anchored at least significant bit end. */ > if (gdbarch_byte_order (objfile_gdbarch) == BFD_ENDIAN_BIG) > - source_offset_bits += obj_size - p->size; > + source_offset_bits += obj_size - (p->offset + p->size); > + else > + source_offset_bits += p->offset; > copy_bitwise (contents, dest_offset_bits, > value_contents_all (p->v.value), > source_offset_bits, > @@ -1888,6 +1893,7 @@ read_pieced_value (struct value *v) > size_t n = this_size_bits; > > /* Cut off at the end of the implicit value. */ > + source_offset_bits += p->offset; > if (source_offset_bits >= obj_size) > break; > if (n > obj_size - source_offset_bits) > @@ -1978,11 +1984,13 @@ write_pieced_value (struct value *to, struct > value *from) > ULONGEST reg_bits = 8 * register_size (arch, gdb_regnum); > > if (gdbarch_byte_order (arch) == BFD_ENDIAN_BIG > - && p->size <= reg_bits) > + && p->offset + p->size < reg_bits) > { > /* Big-endian, and we want less than full size. */ > - dest_offset_bits += reg_bits - p->size; > + dest_offset_bits += reg_bits - (p->offset + p->size); > } > + else > + dest_offset_bits += p->offset; > this_size = bits_to_bytes (dest_offset_bits, this_size_bits); > buffer.reserve (this_size); > > @@ -2019,6 +2027,8 @@ write_pieced_value (struct value *to, struct > value *from) > break; > case DWARF_VALUE_MEMORY: > { > + dest_offset_bits += p->offset; > + > CORE_ADDR start_addr = p->v.mem.addr + dest_offset_bits / 8; > > if (dest_offset_bits % 8 == 0 && this_size_bits % 8 == 0 > diff --git a/gdb/testsuite/gdb.dwarf2/var-access.exp > b/gdb/testsuite/gdb.dwarf2/var-access.exp > index 4787dfb..29d61f8 100644 > --- a/gdb/testsuite/gdb.dwarf2/var-access.exp > +++ b/gdb/testsuite/gdb.dwarf2/var-access.exp > @@ -204,6 +204,22 @@ Dwarf::assemble $asm_file { > piece 1 > } SPECIAL_expr} > } > + # One piece per bitfield. Use piece offsets. > + DW_TAG_variable { > + {name "t3"} > + {type :$struct_t_label} > + {location { > + implicit_value 0x12 0x34 0x56 0x78 0x9a > + bit_piece 32 4 > + const2s -280 > + stack_value > + bit_piece 9 2 > + regx [lindex $dwarf_regnum 0] > + bit_piece 13 14 > + addr $buf_var > + bit_piece 10 42 > + } SPECIAL_expr} > + } > } > } > } > @@ -276,3 +292,24 @@ gdb_test_no_output "set var t2.y = 2641" > gdb_test_no_output "set var t2.z = -400" > gdb_test_no_output "set var t2.x = 200" > gdb_test "print t2.x + t2.y + t2.z" " = 2441" > + > +# Bitfield access through pieces with nonzero piece offsets. > +gdb_test_no_output "set var \$[lindex $regname 0] = 0xa8000" \ > + "init reg for t3.y" > +gdb_test_no_output "set var *(char \[2\] *) (a + 5) = { 70, 82 }" \ > + "init mem for t3.z" > +switch $endian { > + little {set val "u = -1484430527, x = -70, y = 42, z = 145"} > + big {set val "u = 591751049, x = -70, y = 42, z = 101"} > +} Just a nit, I think this (at least the part about u) would be slightly easier to understand if the value was printed in hex, like so (for little endian): u = 0xa7856341, x = 0xffffffba, y = 0x2a, z = 0x91 since we can recognize the nibbles from the implicit_value up there. > +gdb_test "print t3" " = \\{$val\\}" \ > + "initialized t3 from reg and mem" > +gdb_test_no_output "set var t3.y = -1" \ > + "overwrite t3.y" > +gdb_test "print/x \$[lindex $regname 0]" " = 0x7ffc000" \ > + "verify t3.y through reg" > +gdb_test_no_output "set var t3.z = -614" \ > + "overwrite t3.z" > +switch $endian {big {set val "0x59, 0xa2"} little {set val "0x6a, > 0x56"}} > +gdb_test "print/x *(char \[2\] *) (a + 5)" " = \\{$val\\}" \ > + "verify t3.z through mem"