* [patch] workaround gcc46: prologue skip skips too far (PR 12435) #2
@ 2011-07-22 23:04 Jan Kratochvil
2011-07-24 18:16 ` Mark Kettenis
` (2 more replies)
0 siblings, 3 replies; 8+ messages in thread
From: Jan Kratochvil @ 2011-07-22 23:04 UTC (permalink / raw)
To: gdb-patches
Hi,
this is an improved patch of a former:
[patch] workaround gcc46: prologue skip skips too far (PR 12435)
http://sourceware.org/ml/gdb-patches/2011-03/msg01108.html
cancel/FYI: Re: [patch] workaround gcc46: prologue skip skips too far (PR 12435)
http://sourceware.org/ml/gdb-patches/2011-03/msg01123.html
For example `break error' does not work for debugging GDB with gcc-4.6.x.
As gcc-4.6.0 and now even 4.6.1 still has this bug and I have seen a user(s?)
on non-Fedora platform asking about this bug and as there may be enough
binaries out there (although it affects only -O0 -g compilation) coded it
properly I have coded the workaround properly this time.
It does not solve overlays well, but the code just does not work for overlays,
it has no other negative effect.
I will update the code after FSF gcc gets fixed to minimize the workaround.
No regressions on {x86_64,x86_64-m32,i686}-fedora16pre-linux-gnu.
I would welcome a comment whether it is suitable for FSF GDB.
Thanks,
Jan
gdb/
2011-07-22 Jan Kratochvil <jan.kratochvil@redhat.com>
PR breakpoints/12435
* amd64-tdep.c (amd64_skip_prologue): New variables start_pc_sal,
next_sal, buf, offset and xmmreg. Advance PC if it sees the PR.
* dwarf2read.c (process_full_comp_unit): Initialize
amd64_prologue_line_bug.
* symtab.h (struct symtab): New field amd64_prologue_line_bug.
gdb/testsuite/
2011-07-22 Jan Kratochvil <jan.kratochvil@redhat.com>
PR breakpoints/12435
* gdb.arch/amd64-prologue-xmm.c: New file.
* gdb.arch/amd64-prologue-xmm.exp: New file.
* gdb.arch/amd64-prologue-xmm.s: New file.
--- a/gdb/amd64-tdep.c
+++ b/gdb/amd64-tdep.c
@@ -1902,6 +1902,9 @@ amd64_skip_prologue (struct gdbarch *gdbarch, CORE_ADDR start_pc)
{
struct amd64_frame_cache cache;
CORE_ADDR pc;
+ struct symtab_and_line start_pc_sal, next_sal;
+ gdb_byte buf[4 + 8 * 7];
+ int offset, xmmreg;
amd64_init_frame_cache (&cache);
pc = amd64_analyze_prologue (gdbarch, start_pc, 0xffffffffffffffffLL,
@@ -1909,7 +1912,71 @@ amd64_skip_prologue (struct gdbarch *gdbarch, CORE_ADDR start_pc)
if (cache.frameless_p)
return start_pc;
- return pc;
+ /* GCC PR debug/48827 produced false prologue end:
+ 84 c0 test %al,%al
+ 74 23 je after
+ <-- here is 0 lines advance - the false prologue end marker.
+ 0f 29 85 70 ff ff ff movaps %xmm0,-0x90(%rbp)
+ 0f 29 4d 80 movaps %xmm1,-0x80(%rbp)
+ 0f 29 55 90 movaps %xmm2,-0x70(%rbp)
+ 0f 29 5d a0 movaps %xmm3,-0x60(%rbp)
+ 0f 29 65 b0 movaps %xmm4,-0x50(%rbp)
+ 0f 29 6d c0 movaps %xmm5,-0x40(%rbp)
+ 0f 29 75 d0 movaps %xmm6,-0x30(%rbp)
+ 0f 29 7d e0 movaps %xmm7,-0x20(%rbp)
+ after: */
+
+ if (pc == start_pc)
+ return pc;
+
+ start_pc_sal = find_pc_sect_line (start_pc, NULL, 0);
+ if (start_pc_sal.symtab == NULL
+ || !start_pc_sal.symtab->amd64_prologue_line_bug
+ || start_pc_sal.pc != start_pc || pc >= start_pc_sal.end)
+ return pc;
+
+ next_sal = find_pc_sect_line (start_pc_sal.end, NULL, 0);
+ if (next_sal.line != start_pc_sal.line)
+ return pc;
+
+ /* START_PC can be from overlayed memory, ignored here. */
+ if (target_read_memory (next_sal.pc - 4, buf, sizeof (buf)) != 0)
+ return pc;
+
+ /* test %al,%al */
+ if (buf[0] != 0x84 || buf[1] != 0xc0)
+ return pc;
+ /* je AFTER */
+ if (buf[2] != 0x74)
+ return pc;
+
+ offset = 4;
+ for (xmmreg = 0; xmmreg < 8; xmmreg++)
+ {
+ /* movaps %xmmreg?,-0x??(%rbp) */
+ if (buf[offset] != 0x0f || buf[offset + 1] != 0x29
+ || (buf[offset + 2] & 0b00111111) != (xmmreg << 3 | 0b101))
+ return pc;
+
+ if ((buf[offset + 2] & 0b11000000) == 0b01000000)
+ {
+ /* 8-bit displacement. */
+ offset += 4;
+ }
+ else if ((buf[offset + 2] & 0b11000000) == 0b10000000)
+ {
+ /* 32-bit displacement. */
+ offset += 7;
+ }
+ else
+ return pc;
+ }
+
+ /* je AFTER */
+ if (offset - 4 != buf[3])
+ return pc;
+
+ return next_sal.end;
}
\f
--- a/gdb/dwarf2read.c
+++ b/gdb/dwarf2read.c
@@ -4818,6 +4818,9 @@ process_full_comp_unit (struct dwarf2_per_cu_data *per_cu)
if (gcc_4_minor >= 5)
symtab->epilogue_unwind_valid = 1;
+
+ if (gcc_4_minor >= 6)
+ symtab->amd64_prologue_line_bug = 1;
}
if (dwarf2_per_objfile->using_index)
--- a/gdb/symtab.h
+++ b/gdb/symtab.h
@@ -784,6 +784,11 @@ struct symtab
unsigned int epilogue_unwind_valid : 1;
+ /* At least GCC 4.6.0 and 4.6.1 can produce invalid false prologue and marker
+ on amd64. This flag is set independently of the symtab arch. */
+
+ unsigned amd64_prologue_line_bug : 1;
+
/* The macro table for this symtab. Like the blockvector, this
may be shared between different symtabs --- and normally is for
all the symtabs in a given compilation unit. */
--- /dev/null
+++ b/gdb/testsuite/gdb.arch/amd64-prologue-xmm.c
@@ -0,0 +1,38 @@
+/* This testcase is part of GDB, the GNU debugger.
+
+ Copyright 2011 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 3 of the License, or
+ (at your option) any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program. If not, see <http://www.gnu.org/licenses/>. */
+
+static volatile int v, fail;
+
+static void
+func (int i, ...)
+{
+ v = i;
+}
+
+static void
+marker (void)
+{
+}
+
+int
+main (void)
+{
+ func (1);
+ fail = 1;
+ marker ();
+ return 0;
+}
--- /dev/null
+++ b/gdb/testsuite/gdb.arch/amd64-prologue-xmm.exp
@@ -0,0 +1,46 @@
+# Copyright 2011 Free Software Foundation, Inc.
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program. If not, see <http://www.gnu.org/licenses/>.
+
+# Test GCC PR debug/48827 workaround in GDB.
+
+set testfile "amd64-prologue-xmm"
+set srcfile ${testfile}.s
+set csrcfile ${testfile}.c
+set binfile ${objdir}/${subdir}/${testfile}.x
+set opts {}
+
+if [info exists COMPILE] {
+ # make check RUNTESTFLAGS='gdb.arch/amd64-prologue-xmm.exp COMPILE=1'
+ set srcfile ${csrcfile}
+ lappend opts debug optimize=-O0
+} elseif { ![istarget x86_64-*-* ] || ![is_lp64_target] } {
+ verbose "Skipping amd64-prologue-xmm test."
+ return 0
+}
+
+if {[prepare_for_testing ${testfile}.exp ${testfile} $srcfile $opts]} {
+ return -1
+}
+
+if ![runto_main] {
+ return -1
+}
+
+gdb_breakpoint "func"
+gdb_breakpoint "marker"
+
+gdb_continue_to_breakpoint "func"
+
+gdb_test "p fail" " = 0" "stopped at func"
--- /dev/null
+++ b/gdb/testsuite/gdb.arch/amd64-prologue-xmm.s
@@ -0,0 +1,400 @@
+/* This testcase is part of GDB, the GNU debugger.
+
+ Copyright 2011 Free Software Foundation, Inc.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 3 of the License, or
+ (at your option) any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program. If not, see <http://www.gnu.org/licenses/>. */
+
+/* This file is compiled from gdb.arch/amd64-prologue-xmm.c
+ using -g -dA -S. */
+
+ .file "amd64-prologue-xmm.c"
+ .text
+.Ltext0:
+ .local v
+ .comm v,4,4
+ .local fail
+ .comm fail,4,4
+ .type func, @function
+func:
+.LFB0:
+ .file 1 "gdb.arch/amd64-prologue-xmm.c"
+ # gdb.arch/amd64-prologue-xmm.c:22
+ .loc 1 22 0
+ .cfi_startproc
+ # basic block 2
+ pushq %rbp
+ .cfi_def_cfa_offset 16
+ .cfi_offset 6, -16
+ movq %rsp, %rbp
+ .cfi_def_cfa_register 6
+ subq $72, %rsp
+ movq %rsi, -168(%rbp)
+ movq %rdx, -160(%rbp)
+ movq %rcx, -152(%rbp)
+ movq %r8, -144(%rbp)
+ movq %r9, -136(%rbp)
+ testb %al, %al
+ je .L2
+ # basic block 3
+ # gdb.arch/amd64-prologue-xmm.c:22
+ .loc 1 22 0
+ movaps %xmm0, -128(%rbp)
+ movaps %xmm1, -112(%rbp)
+ movaps %xmm2, -96(%rbp)
+ movaps %xmm3, -80(%rbp)
+ movaps %xmm4, -64(%rbp)
+ movaps %xmm5, -48(%rbp)
+ movaps %xmm6, -32(%rbp)
+ movaps %xmm7, -16(%rbp)
+.L2:
+ # basic block 4
+ movl %edi, -180(%rbp)
+ # gdb.arch/amd64-prologue-xmm.c:23
+ .loc 1 23 0
+ movl -180(%rbp), %eax
+ movl %eax, v(%rip)
+ # gdb.arch/amd64-prologue-xmm.c:24
+ .loc 1 24 0
+ leave
+ .cfi_def_cfa 7, 8
+ ret
+ .cfi_endproc
+.LFE0:
+ .size func, .-func
+ .type marker, @function
+marker:
+.LFB1:
+ # gdb.arch/amd64-prologue-xmm.c:28
+ .loc 1 28 0
+ .cfi_startproc
+ # basic block 2
+ pushq %rbp
+ .cfi_def_cfa_offset 16
+ .cfi_offset 6, -16
+ movq %rsp, %rbp
+ .cfi_def_cfa_register 6
+ # gdb.arch/amd64-prologue-xmm.c:29
+ .loc 1 29 0
+ popq %rbp
+ .cfi_def_cfa 7, 8
+ ret
+ .cfi_endproc
+.LFE1:
+ .size marker, .-marker
+ .globl main
+ .type main, @function
+main:
+.LFB2:
+ # gdb.arch/amd64-prologue-xmm.c:33
+ .loc 1 33 0
+ .cfi_startproc
+ # basic block 2
+ pushq %rbp
+ .cfi_def_cfa_offset 16
+ .cfi_offset 6, -16
+ movq %rsp, %rbp
+ .cfi_def_cfa_register 6
+ # gdb.arch/amd64-prologue-xmm.c:34
+ .loc 1 34 0
+ movl $1, %edi
+ movl $0, %eax
+ call func
+ # gdb.arch/amd64-prologue-xmm.c:35
+ .loc 1 35 0
+ movl $1, fail(%rip)
+ # gdb.arch/amd64-prologue-xmm.c:36
+ .loc 1 36 0
+ call marker
+ # gdb.arch/amd64-prologue-xmm.c:37
+ .loc 1 37 0
+ movl $0, %eax
+ # gdb.arch/amd64-prologue-xmm.c:38
+ .loc 1 38 0
+ popq %rbp
+ .cfi_def_cfa 7, 8
+ ret
+ .cfi_endproc
+.LFE2:
+ .size main, .-main
+.Letext0:
+ .section .debug_info,"",@progbits
+.Ldebug_info0:
+ .long 0xc0 # Length of Compilation Unit Info
+ .value 0x4 # DWARF version number
+ .long .Ldebug_abbrev0 # Offset Into Abbrev. Section
+ .byte 0x8 # Pointer Size (in bytes)
+ .uleb128 0x1 # (DIE (0xb) DW_TAG_compile_unit)
+ .long .LASF1 # DW_AT_producer: "GNU C 4.6.1 20110715 (Red Hat 4.6.1-3)"
+ .byte 0x1 # DW_AT_language
+ .long .LASF2 # DW_AT_name: "gdb.arch/amd64-prologue-xmm.c"
+ .long .LASF3 # DW_AT_comp_dir: ""
+ .quad .Ltext0 # DW_AT_low_pc
+ .quad .Letext0 # DW_AT_high_pc
+ .long .Ldebug_line0 # DW_AT_stmt_list
+ .uleb128 0x2 # (DIE (0x2d) DW_TAG_subprogram)
+ .long .LASF4 # DW_AT_name: "func"
+ .byte 0x1 # DW_AT_decl_file (gdb.arch/amd64-prologue-xmm.c)
+ .byte 0x15 # DW_AT_decl_line
+ # DW_AT_prototyped
+ .quad .LFB0 # DW_AT_low_pc
+ .quad .LFE0 # DW_AT_high_pc
+ .uleb128 0x1 # DW_AT_frame_base
+ .byte 0x9c # DW_OP_call_frame_cfa
+ # DW_AT_GNU_all_call_sites
+ .long 0x59 # DW_AT_sibling
+ .uleb128 0x3 # (DIE (0x4a) DW_TAG_formal_parameter)
+ .ascii "i\0" # DW_AT_name
+ .byte 0x1 # DW_AT_decl_file (gdb.arch/amd64-prologue-xmm.c)
+ .byte 0x15 # DW_AT_decl_line
+ .long 0x59 # DW_AT_type
+ .uleb128 0x3 # DW_AT_location
+ .byte 0x91 # DW_OP_fbreg
+ .sleb128 -196
+ .uleb128 0x4 # (DIE (0x57) DW_TAG_unspecified_parameters)
+ .byte 0 # end of children of DIE 0x2d
+ .uleb128 0x5 # (DIE (0x59) DW_TAG_base_type)
+ .byte 0x4 # DW_AT_byte_size
+ .byte 0x5 # DW_AT_encoding
+ .ascii "int\0" # DW_AT_name
+ .uleb128 0x6 # (DIE (0x60) DW_TAG_subprogram)
+ .long .LASF5 # DW_AT_name: "marker"
+ .byte 0x1 # DW_AT_decl_file (gdb.arch/amd64-prologue-xmm.c)
+ .byte 0x1b # DW_AT_decl_line
+ # DW_AT_prototyped
+ .quad .LFB1 # DW_AT_low_pc
+ .quad .LFE1 # DW_AT_high_pc
+ .uleb128 0x1 # DW_AT_frame_base
+ .byte 0x9c # DW_OP_call_frame_cfa
+ # DW_AT_GNU_all_call_sites
+ .uleb128 0x7 # (DIE (0x79) DW_TAG_subprogram)
+ # DW_AT_external
+ .long .LASF6 # DW_AT_name: "main"
+ .byte 0x1 # DW_AT_decl_file (gdb.arch/amd64-prologue-xmm.c)
+ .byte 0x20 # DW_AT_decl_line
+ # DW_AT_prototyped
+ .long 0x59 # DW_AT_type
+ .quad .LFB2 # DW_AT_low_pc
+ .quad .LFE2 # DW_AT_high_pc
+ .uleb128 0x1 # DW_AT_frame_base
+ .byte 0x9c # DW_OP_call_frame_cfa
+ # DW_AT_GNU_all_tail_call_sites
+ .uleb128 0x8 # (DIE (0x96) DW_TAG_variable)
+ .ascii "v\0" # DW_AT_name
+ .byte 0x1 # DW_AT_decl_file (gdb.arch/amd64-prologue-xmm.c)
+ .byte 0x12 # DW_AT_decl_line
+ .long 0xa9 # DW_AT_type
+ .uleb128 0x9 # DW_AT_location
+ .byte 0x3 # DW_OP_addr
+ .quad v
+ .uleb128 0x9 # (DIE (0xa9) DW_TAG_volatile_type)
+ .long 0x59 # DW_AT_type
+ .uleb128 0xa # (DIE (0xae) DW_TAG_variable)
+ .long .LASF0 # DW_AT_name: "fail"
+ .byte 0x1 # DW_AT_decl_file (gdb.arch/amd64-prologue-xmm.c)
+ .byte 0x12 # DW_AT_decl_line
+ .long 0xa9 # DW_AT_type
+ .uleb128 0x9 # DW_AT_location
+ .byte 0x3 # DW_OP_addr
+ .quad fail
+ .byte 0 # end of children of DIE 0xb
+ .section .debug_abbrev,"",@progbits
+.Ldebug_abbrev0:
+ .uleb128 0x1 # (abbrev code)
+ .uleb128 0x11 # (TAG: DW_TAG_compile_unit)
+ .byte 0x1 # DW_children_yes
+ .uleb128 0x25 # (DW_AT_producer)
+ .uleb128 0xe # (DW_FORM_strp)
+ .uleb128 0x13 # (DW_AT_language)
+ .uleb128 0xb # (DW_FORM_data1)
+ .uleb128 0x3 # (DW_AT_name)
+ .uleb128 0xe # (DW_FORM_strp)
+ .uleb128 0x1b # (DW_AT_comp_dir)
+ .uleb128 0xe # (DW_FORM_strp)
+ .uleb128 0x11 # (DW_AT_low_pc)
+ .uleb128 0x1 # (DW_FORM_addr)
+ .uleb128 0x12 # (DW_AT_high_pc)
+ .uleb128 0x1 # (DW_FORM_addr)
+ .uleb128 0x10 # (DW_AT_stmt_list)
+ .uleb128 0x17 # (DW_FORM_sec_offset)
+ .byte 0
+ .byte 0
+ .uleb128 0x2 # (abbrev code)
+ .uleb128 0x2e # (TAG: DW_TAG_subprogram)
+ .byte 0x1 # DW_children_yes
+ .uleb128 0x3 # (DW_AT_name)
+ .uleb128 0xe # (DW_FORM_strp)
+ .uleb128 0x3a # (DW_AT_decl_file)
+ .uleb128 0xb # (DW_FORM_data1)
+ .uleb128 0x3b # (DW_AT_decl_line)
+ .uleb128 0xb # (DW_FORM_data1)
+ .uleb128 0x27 # (DW_AT_prototyped)
+ .uleb128 0x19 # (DW_FORM_flag_present)
+ .uleb128 0x11 # (DW_AT_low_pc)
+ .uleb128 0x1 # (DW_FORM_addr)
+ .uleb128 0x12 # (DW_AT_high_pc)
+ .uleb128 0x1 # (DW_FORM_addr)
+ .uleb128 0x40 # (DW_AT_frame_base)
+ .uleb128 0x18 # (DW_FORM_exprloc)
+ .uleb128 0x2117 # (DW_AT_GNU_all_call_sites)
+ .uleb128 0x19 # (DW_FORM_flag_present)
+ .uleb128 0x1 # (DW_AT_sibling)
+ .uleb128 0x13 # (DW_FORM_ref4)
+ .byte 0
+ .byte 0
+ .uleb128 0x3 # (abbrev code)
+ .uleb128 0x5 # (TAG: DW_TAG_formal_parameter)
+ .byte 0 # DW_children_no
+ .uleb128 0x3 # (DW_AT_name)
+ .uleb128 0x8 # (DW_FORM_string)
+ .uleb128 0x3a # (DW_AT_decl_file)
+ .uleb128 0xb # (DW_FORM_data1)
+ .uleb128 0x3b # (DW_AT_decl_line)
+ .uleb128 0xb # (DW_FORM_data1)
+ .uleb128 0x49 # (DW_AT_type)
+ .uleb128 0x13 # (DW_FORM_ref4)
+ .uleb128 0x2 # (DW_AT_location)
+ .uleb128 0x18 # (DW_FORM_exprloc)
+ .byte 0
+ .byte 0
+ .uleb128 0x4 # (abbrev code)
+ .uleb128 0x18 # (TAG: DW_TAG_unspecified_parameters)
+ .byte 0 # DW_children_no
+ .byte 0
+ .byte 0
+ .uleb128 0x5 # (abbrev code)
+ .uleb128 0x24 # (TAG: DW_TAG_base_type)
+ .byte 0 # DW_children_no
+ .uleb128 0xb # (DW_AT_byte_size)
+ .uleb128 0xb # (DW_FORM_data1)
+ .uleb128 0x3e # (DW_AT_encoding)
+ .uleb128 0xb # (DW_FORM_data1)
+ .uleb128 0x3 # (DW_AT_name)
+ .uleb128 0x8 # (DW_FORM_string)
+ .byte 0
+ .byte 0
+ .uleb128 0x6 # (abbrev code)
+ .uleb128 0x2e # (TAG: DW_TAG_subprogram)
+ .byte 0 # DW_children_no
+ .uleb128 0x3 # (DW_AT_name)
+ .uleb128 0xe # (DW_FORM_strp)
+ .uleb128 0x3a # (DW_AT_decl_file)
+ .uleb128 0xb # (DW_FORM_data1)
+ .uleb128 0x3b # (DW_AT_decl_line)
+ .uleb128 0xb # (DW_FORM_data1)
+ .uleb128 0x27 # (DW_AT_prototyped)
+ .uleb128 0x19 # (DW_FORM_flag_present)
+ .uleb128 0x11 # (DW_AT_low_pc)
+ .uleb128 0x1 # (DW_FORM_addr)
+ .uleb128 0x12 # (DW_AT_high_pc)
+ .uleb128 0x1 # (DW_FORM_addr)
+ .uleb128 0x40 # (DW_AT_frame_base)
+ .uleb128 0x18 # (DW_FORM_exprloc)
+ .uleb128 0x2117 # (DW_AT_GNU_all_call_sites)
+ .uleb128 0x19 # (DW_FORM_flag_present)
+ .byte 0
+ .byte 0
+ .uleb128 0x7 # (abbrev code)
+ .uleb128 0x2e # (TAG: DW_TAG_subprogram)
+ .byte 0 # DW_children_no
+ .uleb128 0x3f # (DW_AT_external)
+ .uleb128 0x19 # (DW_FORM_flag_present)
+ .uleb128 0x3 # (DW_AT_name)
+ .uleb128 0xe # (DW_FORM_strp)
+ .uleb128 0x3a # (DW_AT_decl_file)
+ .uleb128 0xb # (DW_FORM_data1)
+ .uleb128 0x3b # (DW_AT_decl_line)
+ .uleb128 0xb # (DW_FORM_data1)
+ .uleb128 0x27 # (DW_AT_prototyped)
+ .uleb128 0x19 # (DW_FORM_flag_present)
+ .uleb128 0x49 # (DW_AT_type)
+ .uleb128 0x13 # (DW_FORM_ref4)
+ .uleb128 0x11 # (DW_AT_low_pc)
+ .uleb128 0x1 # (DW_FORM_addr)
+ .uleb128 0x12 # (DW_AT_high_pc)
+ .uleb128 0x1 # (DW_FORM_addr)
+ .uleb128 0x40 # (DW_AT_frame_base)
+ .uleb128 0x18 # (DW_FORM_exprloc)
+ .uleb128 0x2116 # (DW_AT_GNU_all_tail_call_sites)
+ .uleb128 0x19 # (DW_FORM_flag_present)
+ .byte 0
+ .byte 0
+ .uleb128 0x8 # (abbrev code)
+ .uleb128 0x34 # (TAG: DW_TAG_variable)
+ .byte 0 # DW_children_no
+ .uleb128 0x3 # (DW_AT_name)
+ .uleb128 0x8 # (DW_FORM_string)
+ .uleb128 0x3a # (DW_AT_decl_file)
+ .uleb128 0xb # (DW_FORM_data1)
+ .uleb128 0x3b # (DW_AT_decl_line)
+ .uleb128 0xb # (DW_FORM_data1)
+ .uleb128 0x49 # (DW_AT_type)
+ .uleb128 0x13 # (DW_FORM_ref4)
+ .uleb128 0x2 # (DW_AT_location)
+ .uleb128 0x18 # (DW_FORM_exprloc)
+ .byte 0
+ .byte 0
+ .uleb128 0x9 # (abbrev code)
+ .uleb128 0x35 # (TAG: DW_TAG_volatile_type)
+ .byte 0 # DW_children_no
+ .uleb128 0x49 # (DW_AT_type)
+ .uleb128 0x13 # (DW_FORM_ref4)
+ .byte 0
+ .byte 0
+ .uleb128 0xa # (abbrev code)
+ .uleb128 0x34 # (TAG: DW_TAG_variable)
+ .byte 0 # DW_children_no
+ .uleb128 0x3 # (DW_AT_name)
+ .uleb128 0xe # (DW_FORM_strp)
+ .uleb128 0x3a # (DW_AT_decl_file)
+ .uleb128 0xb # (DW_FORM_data1)
+ .uleb128 0x3b # (DW_AT_decl_line)
+ .uleb128 0xb # (DW_FORM_data1)
+ .uleb128 0x49 # (DW_AT_type)
+ .uleb128 0x13 # (DW_FORM_ref4)
+ .uleb128 0x2 # (DW_AT_location)
+ .uleb128 0x18 # (DW_FORM_exprloc)
+ .byte 0
+ .byte 0
+ .byte 0
+ .section .debug_aranges,"",@progbits
+ .long 0x2c # Length of Address Ranges Info
+ .value 0x2 # DWARF Version
+ .long .Ldebug_info0 # Offset of Compilation Unit Info
+ .byte 0x8 # Size of Address
+ .byte 0 # Size of Segment Descriptor
+ .value 0 # Pad to 16 byte boundary
+ .value 0
+ .quad .Ltext0 # Address
+ .quad .Letext0-.Ltext0 # Length
+ .quad 0
+ .quad 0
+ .section .debug_line,"",@progbits
+.Ldebug_line0:
+ .section .debug_str,"MS",@progbits,1
+.LASF3:
+ .string ""
+.LASF0:
+ .string "fail"
+.LASF4:
+ .string "func"
+.LASF1:
+ .string "GNU C 4.6.1 20110715 (Red Hat 4.6.1-3)"
+.LASF2:
+ .string "gdb.arch/amd64-prologue-xmm.c"
+.LASF5:
+ .string "marker"
+.LASF6:
+ .string "main"
+ .ident "GCC: (GNU) 4.6.1 20110715 (Red Hat 4.6.1-3)"
+ .section .note.GNU-stack,"",@progbits
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [patch] workaround gcc46: prologue skip skips too far (PR 12435) #2
2011-07-22 23:04 [patch] workaround gcc46: prologue skip skips too far (PR 12435) #2 Jan Kratochvil
@ 2011-07-24 18:16 ` Mark Kettenis
2011-07-24 20:00 ` Jan Kratochvil
2011-09-08 15:54 ` Jan Kratochvil
2011-09-08 17:31 ` Doug Evans
2 siblings, 1 reply; 8+ messages in thread
From: Mark Kettenis @ 2011-07-24 18:16 UTC (permalink / raw)
To: jan.kratochvil; +Cc: gdb-patches
> Date: Fri, 22 Jul 2011 23:58:21 +0200
> From: Jan Kratochvil <jan.kratochvil@redhat.com>
>
> Hi,
>
> this is an improved patch of a former:
> [patch] workaround gcc46: prologue skip skips too far (PR 12435)
> http://sourceware.org/ml/gdb-patches/2011-03/msg01108.html
> cancel/FYI: Re: [patch] workaround gcc46: prologue skip skips too far (PR 12435)
> http://sourceware.org/ml/gdb-patches/2011-03/msg01123.html
>
> For example `break error' does not work for debugging GDB with gcc-4.6.x.
>
> As gcc-4.6.0 and now even 4.6.1 still has this bug and I have seen a user(s?)
> on non-Fedora platform asking about this bug and as there may be enough
> binaries out there (although it affects only -O0 -g compilation) coded it
> properly I have coded the workaround properly this time.
>
> It does not solve overlays well, but the code just does not work for
> overlays, it has no other negative effect.
Sorry, but what do you mean with "overlays" here?
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [patch] workaround gcc46: prologue skip skips too far (PR 12435) #2
2011-07-24 18:16 ` Mark Kettenis
@ 2011-07-24 20:00 ` Jan Kratochvil
0 siblings, 0 replies; 8+ messages in thread
From: Jan Kratochvil @ 2011-07-24 20:00 UTC (permalink / raw)
To: Mark Kettenis; +Cc: gdb-patches
On Sun, 24 Jul 2011 20:03:53 +0200, Mark Kettenis wrote:
> > It does not solve overlays well, but the code just does not work for
> > overlays, it has no other negative effect.
>
> Sorry, but what do you mean with "overlays" here?
See skip_prologue_sal:
/* If the function is in an unmapped overlay, use its unmapped LMA address,
so that gdbarch_skip_prologue has something unique to work on. */
if (section_is_overlay (section) && !section_is_mapped (section))
pc = overlay_unmapped_address (pc, section);
/* Skip "first line" of function (which is actually its prologue). */
pc += gdbarch_deprecated_function_start_offset (gdbarch);
if (skip)
pc = gdbarch_skip_prologue (gdbarch, pc);
/* For overlays, map pc back into its mapped VMA range. */
pc = overlay_mapped_address (pc, section);
So in gdbarch_skip_prologue the PC is LMA, I do not know how
gdbarch_skip_prologue could read the bytes at PC (without introducing new
parameter(s) to gdbarch_skip_prologue).
Thanks,
Jan
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [patch] workaround gcc46: prologue skip skips too far (PR 12435) #2
2011-07-22 23:04 [patch] workaround gcc46: prologue skip skips too far (PR 12435) #2 Jan Kratochvil
2011-07-24 18:16 ` Mark Kettenis
@ 2011-09-08 15:54 ` Jan Kratochvil
2011-09-08 17:31 ` Doug Evans
2 siblings, 0 replies; 8+ messages in thread
From: Jan Kratochvil @ 2011-09-08 15:54 UTC (permalink / raw)
To: gdb-patches
On Fri, 22 Jul 2011 23:58:21 +0200, Jan Kratochvil wrote:
> gdb/
> 2011-07-22 Jan Kratochvil <jan.kratochvil@redhat.com>
>
> PR breakpoints/12435
> * amd64-tdep.c (amd64_skip_prologue): New variables start_pc_sal,
> next_sal, buf, offset and xmmreg. Advance PC if it sees the PR.
> * dwarf2read.c (process_full_comp_unit): Initialize
> amd64_prologue_line_bug.
> * symtab.h (struct symtab): New field amd64_prologue_line_bug.
>
> gdb/testsuite/
> 2011-07-22 Jan Kratochvil <jan.kratochvil@redhat.com>
>
> PR breakpoints/12435
> * gdb.arch/amd64-prologue-xmm.c: New file.
> * gdb.arch/amd64-prologue-xmm.exp: New file.
> * gdb.arch/amd64-prologue-xmm.s: New file.
Checked in:
http://sourceware.org/ml/gdb-cvs/2011-09/msg00046.html
Thanks,
Jan
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [patch] workaround gcc46: prologue skip skips too far (PR 12435) #2
2011-07-22 23:04 [patch] workaround gcc46: prologue skip skips too far (PR 12435) #2 Jan Kratochvil
2011-07-24 18:16 ` Mark Kettenis
2011-09-08 15:54 ` Jan Kratochvil
@ 2011-09-08 17:31 ` Doug Evans
2011-09-08 19:13 ` Jan Kratochvil
2 siblings, 1 reply; 8+ messages in thread
From: Doug Evans @ 2011-09-08 17:31 UTC (permalink / raw)
To: Jan Kratochvil; +Cc: gdb-patches
On Fri, Jul 22, 2011 at 2:58 PM, Jan Kratochvil
<jan.kratochvil@redhat.com> wrote:
> Hi,
>
> this is an improved patch of a former:
> [patch] workaround gcc46: prologue skip skips too far (PR 12435)
> http://sourceware.org/ml/gdb-patches/2011-03/msg01108.html
> cancel/FYI: Re: [patch] workaround gcc46: prologue skip skips too far (PR 12435)
> http://sourceware.org/ml/gdb-patches/2011-03/msg01123.html
>
> [...]
>
> I will update the code after FSF gcc gets fixed to minimize the workaround.
I realize this is checked in, but a couple of comments for maybe the
next patch for this.
> [...]
>
> --- a/gdb/amd64-tdep.c
> +++ b/gdb/amd64-tdep.c
> @@ -1902,6 +1902,9 @@ amd64_skip_prologue (struct gdbarch *gdbarch, CORE_ADDR start_pc)
> {
> struct amd64_frame_cache cache;
> CORE_ADDR pc;
> + struct symtab_and_line start_pc_sal, next_sal;
> + gdb_byte buf[4 + 8 * 7];
> + int offset, xmmreg;
>
> amd64_init_frame_cache (&cache);
> pc = amd64_analyze_prologue (gdbarch, start_pc, 0xffffffffffffffffLL,
> @@ -1909,7 +1912,71 @@ amd64_skip_prologue (struct gdbarch *gdbarch, CORE_ADDR start_pc)
> if (cache.frameless_p)
> return start_pc;
>
> - return pc;
> + /* GCC PR debug/48827 produced false prologue end:
> + 84 c0 test %al,%al
> + 74 23 je after
> + <-- here is 0 lines advance - the false prologue end marker.
> + 0f 29 85 70 ff ff ff movaps %xmm0,-0x90(%rbp)
> + 0f 29 4d 80 movaps %xmm1,-0x80(%rbp)
> + 0f 29 55 90 movaps %xmm2,-0x70(%rbp)
> + 0f 29 5d a0 movaps %xmm3,-0x60(%rbp)
> + 0f 29 65 b0 movaps %xmm4,-0x50(%rbp)
> + 0f 29 6d c0 movaps %xmm5,-0x40(%rbp)
> + 0f 29 75 d0 movaps %xmm6,-0x30(%rbp)
> + 0f 29 7d e0 movaps %xmm7,-0x20(%rbp)
> + after: */
> +
> + if (pc == start_pc)
> + return pc;
> +
> + start_pc_sal = find_pc_sect_line (start_pc, NULL, 0);
> + if (start_pc_sal.symtab == NULL
> + || !start_pc_sal.symtab->amd64_prologue_line_bug
> + || start_pc_sal.pc != start_pc || pc >= start_pc_sal.end)
> + return pc;
> +
> + next_sal = find_pc_sect_line (start_pc_sal.end, NULL, 0);
> + if (next_sal.line != start_pc_sal.line)
> + return pc;
> +
> + /* START_PC can be from overlayed memory, ignored here. */
> + if (target_read_memory (next_sal.pc - 4, buf, sizeof (buf)) != 0)
> + return pc;
> +
> + /* test %al,%al */
> + if (buf[0] != 0x84 || buf[1] != 0xc0)
> + return pc;
> + /* je AFTER */
> + if (buf[2] != 0x74)
> + return pc;
> +
> + offset = 4;
> + for (xmmreg = 0; xmmreg < 8; xmmreg++)
> + {
> + /* movaps %xmmreg?,-0x??(%rbp) */
> + if (buf[offset] != 0x0f || buf[offset + 1] != 0x29
> + || (buf[offset + 2] & 0b00111111) != (xmmreg << 3 | 0b101))
> + return pc;
> +
> + if ((buf[offset + 2] & 0b11000000) == 0b01000000)
> + {
> + /* 8-bit displacement. */
> + offset += 4;
> + }
> + else if ((buf[offset + 2] & 0b11000000) == 0b10000000)
> + {
> + /* 32-bit displacement. */
> + offset += 7;
> + }
> + else
> + return pc;
> + }
> +
> + /* je AFTER */
> + if (offset - 4 != buf[3])
> + return pc;
> +
> + return next_sal.end;
> }
These kinds of functions tend to grow and grow and become hard to maintain.
IWBN if all this new code is about a specific issue then tuck it away
in its own function.
[Or at least document both the start and end of it.]
> [...]
>
> --- a/gdb/symtab.h
> +++ b/gdb/symtab.h
> @@ -784,6 +784,11 @@ struct symtab
>
> unsigned int epilogue_unwind_valid : 1;
>
> + /* At least GCC 4.6.0 and 4.6.1 can produce invalid false prologue and marker
> + on amd64. This flag is set independently of the symtab arch. */
> +
> + unsigned amd64_prologue_line_bug : 1;
> +
> /* The macro table for this symtab. Like the blockvector, this
> may be shared between different symtabs --- and normally is for
> all the symtabs in a given compilation unit. */
I can hear the screams of pushback if a random net person wanted to
check in anything architecture specific to the architecture
independent part of gdb. :-)
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [patch] workaround gcc46: prologue skip skips too far (PR 12435) #2
2011-09-08 17:31 ` Doug Evans
@ 2011-09-08 19:13 ` Jan Kratochvil
2011-09-09 19:52 ` Jan Kratochvil
0 siblings, 1 reply; 8+ messages in thread
From: Jan Kratochvil @ 2011-09-08 19:13 UTC (permalink / raw)
To: Doug Evans; +Cc: gdb-patches
On Thu, 08 Sep 2011 19:15:59 +0200, Doug Evans wrote:
> These kinds of functions tend to grow and grow and become hard to maintain.
> IWBN if all this new code is about a specific issue then tuck it away
> in its own function.
I agree it is a good improvement.
> > + Â /* At least GCC 4.6.0 and 4.6.1 can produce invalid false prologue and marker
> > + Â Â on amd64. Â This flag is set independently of the symtab arch. Â */
> > +
> > + Â unsigned amd64_prologue_line_bug : 1;
>
> I can hear the screams of pushback if a random net person wanted to
> check in anything architecture specific to the architecture
> independent part of gdb. :-)
That is a good comment. There is also symtab->producer, not just cu->producer
so it is not needed at all.
I will check it in.
Thanks,
Jan
gdb/
2011-09-08 Jan Kratochvil <jan.kratochvil@redhat.com>
Code cleanup.
* amd64-tdep.c (amd64_skip_prologue): Move the XMM code to ...
(amd64_skip_xmm_prologue): ... this new function. Describe its
parameters. No longer use amd64_prologue_line_bug.
* defs.h (producer_is_gcc_ge_4): New declaration.
* dwarf2read.c (producer_is_gcc_ge_4): Move to utils.c.
(process_full_comp_unit): Update its caller. Remove
amd64_prologue_line_bug initialization.
* symtab.h (struct symtab): Remove field amd64_prologue_line_bug.
* utils.c (producer_is_gcc_ge_4): Moved here from dwarf2read.c.
--- a/gdb/amd64-tdep.c
+++ b/gdb/amd64-tdep.c
@@ -1910,43 +1910,37 @@ amd64_analyze_prologue (struct gdbarch *gdbarch,
return pc;
}
-/* Return PC of first real instruction. */
+/* Work around false termination of prologue - GCC PR debug/48827.
+
+ START_PC is the first instruction of a function, PC is its minimal already
+ determined advanced address. Function returns PC if it has nothing to do.
+
+ 84 c0 test %al,%al
+ 74 23 je after
+ <-- here is 0 lines advance - the false prologue end marker.
+ 0f 29 85 70 ff ff ff movaps %xmm0,-0x90(%rbp)
+ 0f 29 4d 80 movaps %xmm1,-0x80(%rbp)
+ 0f 29 55 90 movaps %xmm2,-0x70(%rbp)
+ 0f 29 5d a0 movaps %xmm3,-0x60(%rbp)
+ 0f 29 65 b0 movaps %xmm4,-0x50(%rbp)
+ 0f 29 6d c0 movaps %xmm5,-0x40(%rbp)
+ 0f 29 75 d0 movaps %xmm6,-0x30(%rbp)
+ 0f 29 7d e0 movaps %xmm7,-0x20(%rbp)
+ after: */
static CORE_ADDR
-amd64_skip_prologue (struct gdbarch *gdbarch, CORE_ADDR start_pc)
+amd64_skip_xmm_prologue (CORE_ADDR pc, CORE_ADDR start_pc)
{
- struct amd64_frame_cache cache;
- CORE_ADDR pc;
struct symtab_and_line start_pc_sal, next_sal;
gdb_byte buf[4 + 8 * 7];
int offset, xmmreg;
- amd64_init_frame_cache (&cache);
- pc = amd64_analyze_prologue (gdbarch, start_pc, 0xffffffffffffffffLL,
- &cache);
- if (cache.frameless_p)
- return start_pc;
-
- /* GCC PR debug/48827 produced false prologue end:
- 84 c0 test %al,%al
- 74 23 je after
- <-- here is 0 lines advance - the false prologue end marker.
- 0f 29 85 70 ff ff ff movaps %xmm0,-0x90(%rbp)
- 0f 29 4d 80 movaps %xmm1,-0x80(%rbp)
- 0f 29 55 90 movaps %xmm2,-0x70(%rbp)
- 0f 29 5d a0 movaps %xmm3,-0x60(%rbp)
- 0f 29 65 b0 movaps %xmm4,-0x50(%rbp)
- 0f 29 6d c0 movaps %xmm5,-0x40(%rbp)
- 0f 29 75 d0 movaps %xmm6,-0x30(%rbp)
- 0f 29 7d e0 movaps %xmm7,-0x20(%rbp)
- after: */
-
if (pc == start_pc)
return pc;
start_pc_sal = find_pc_sect_line (start_pc, NULL, 0);
if (start_pc_sal.symtab == NULL
- || !start_pc_sal.symtab->amd64_prologue_line_bug
+ || producer_is_gcc_ge_4 (start_pc_sal.symtab->producer) < 6
|| start_pc_sal.pc != start_pc || pc >= start_pc_sal.end)
return pc;
@@ -1993,6 +1987,23 @@ amd64_skip_prologue (struct gdbarch *gdbarch, CORE_ADDR start_pc)
return next_sal.end;
}
+
+/* Return PC of first real instruction. */
+
+static CORE_ADDR
+amd64_skip_prologue (struct gdbarch *gdbarch, CORE_ADDR start_pc)
+{
+ struct amd64_frame_cache cache;
+ CORE_ADDR pc;
+
+ amd64_init_frame_cache (&cache);
+ pc = amd64_analyze_prologue (gdbarch, start_pc, 0xffffffffffffffffLL,
+ &cache);
+ if (cache.frameless_p)
+ return start_pc;
+
+ return amd64_skip_xmm_prologue (pc, start_pc);
+}
\f
/* Normal frames. */
--- a/gdb/defs.h
+++ b/gdb/defs.h
@@ -431,6 +431,8 @@ extern int parse_pid_to_attach (char *args);
extern struct cleanup *make_bpstat_clear_actions_cleanup (void);
+extern int producer_is_gcc_ge_4 (const char *producer);
+
/* From demangle.c */
extern void set_demangling_style (char *);
--- a/gdb/dwarf2read.c
+++ b/gdb/dwarf2read.c
@@ -4723,50 +4723,6 @@ compute_delayed_physnames (struct dwarf2_cu *cu)
}
}
-/* Check for GCC >= 4.x. Return minor version (x) of 4.x in such case. If it
- is not GCC or it is GCC older than 4.x return -1. If it is GCC 5.x or
- higher return INT_MAX. */
-
-static int
-producer_is_gcc_ge_4 (struct dwarf2_cu *cu)
-{
- const char *cs;
- int major, minor;
-
- if (cu->producer == NULL)
- {
- /* For unknown compilers expect their behavior is not compliant. For GCC
- this case can also happen for -gdwarf-4 type units supported since
- gcc-4.5. */
-
- return -1;
- }
-
- /* Skip any identifier after "GNU " - such as "C++" or "Java". */
-
- if (strncmp (cu->producer, "GNU ", strlen ("GNU ")) != 0)
- {
- /* For non-GCC compilers expect their behavior is not compliant. */
-
- return -1;
- }
- cs = &cu->producer[strlen ("GNU ")];
- while (*cs && !isdigit (*cs))
- cs++;
- if (sscanf (cs, "%d.%d", &major, &minor) != 2)
- {
- /* Not recognized as GCC. */
-
- return -1;
- }
-
- if (major < 4)
- return -1;
- if (major > 4)
- return INT_MAX;
- return minor;
-}
-
/* Generate full symbol information for PST and CU, whose DIEs have
already been loaded into memory. */
@@ -4806,7 +4762,7 @@ process_full_comp_unit (struct dwarf2_per_cu_data *per_cu)
if (symtab != NULL)
{
- int gcc_4_minor = producer_is_gcc_ge_4 (cu);
+ int gcc_4_minor = producer_is_gcc_ge_4 (cu->producer);
/* Set symtab language to language from DW_AT_language. If the
compilation is from a C file generated by language preprocessors, do
@@ -4829,9 +4785,6 @@ process_full_comp_unit (struct dwarf2_per_cu_data *per_cu)
if (gcc_4_minor >= 5)
symtab->epilogue_unwind_valid = 1;
-
- if (gcc_4_minor >= 6)
- symtab->amd64_prologue_line_bug = 1;
}
if (dwarf2_per_objfile->using_index)
--- a/gdb/symtab.h
+++ b/gdb/symtab.h
@@ -784,11 +784,6 @@ struct symtab
unsigned int epilogue_unwind_valid : 1;
- /* At least GCC 4.6.0 and 4.6.1 can produce invalid false prologue and marker
- on amd64. This flag is set independently of the symtab arch. */
-
- unsigned amd64_prologue_line_bug : 1;
-
/* The macro table for this symtab. Like the blockvector, this
may be shared between different symtabs --- and normally is for
all the symtabs in a given compilation unit. */
--- a/gdb/utils.c
+++ b/gdb/utils.c
@@ -3691,6 +3691,50 @@ make_bpstat_clear_actions_cleanup (void)
return make_cleanup (do_bpstat_clear_actions_cleanup, NULL);
}
+/* Check for GCC >= 4.x according to the symtab->producer string. Return minor
+ version (x) of 4.x in such case. If it is not GCC or it is GCC older than
+ 4.x return -1. If it is GCC 5.x or higher return INT_MAX. */
+
+int
+producer_is_gcc_ge_4 (const char *producer)
+{
+ const char *cs;
+ int major, minor;
+
+ if (producer == NULL)
+ {
+ /* For unknown compilers expect their behavior is not compliant. For GCC
+ this case can also happen for -gdwarf-4 type units supported since
+ gcc-4.5. */
+
+ return -1;
+ }
+
+ /* Skip any identifier after "GNU " - such as "C++" or "Java". */
+
+ if (strncmp (producer, "GNU ", strlen ("GNU ")) != 0)
+ {
+ /* For non-GCC compilers expect their behavior is not compliant. */
+
+ return -1;
+ }
+ cs = &producer[strlen ("GNU ")];
+ while (*cs && !isdigit (*cs))
+ cs++;
+ if (sscanf (cs, "%d.%d", &major, &minor) != 2)
+ {
+ /* Not recognized as GCC. */
+
+ return -1;
+ }
+
+ if (major < 4)
+ return -1;
+ if (major > 4)
+ return INT_MAX;
+ return minor;
+}
+
/* Provide a prototype to silence -Wmissing-prototypes. */
extern initialize_file_ftype _initialize_utils;
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [patch] workaround gcc46: prologue skip skips too far (PR 12435) #2
2011-09-08 19:13 ` Jan Kratochvil
@ 2011-09-09 19:52 ` Jan Kratochvil
2011-09-12 17:23 ` Doug Evans
0 siblings, 1 reply; 8+ messages in thread
From: Jan Kratochvil @ 2011-09-09 19:52 UTC (permalink / raw)
To: Doug Evans; +Cc: gdb-patches
On Thu, 08 Sep 2011 20:22:27 +0200, Jan Kratochvil wrote:
> gdb/
> 2011-09-08 Jan Kratochvil <jan.kratochvil@redhat.com>
>
> Code cleanup.
> * amd64-tdep.c (amd64_skip_prologue): Move the XMM code to ...
> (amd64_skip_xmm_prologue): ... this new function. Describe its
> parameters. No longer use amd64_prologue_line_bug.
> * defs.h (producer_is_gcc_ge_4): New declaration.
> * dwarf2read.c (producer_is_gcc_ge_4): Move to utils.c.
> (process_full_comp_unit): Update its caller. Remove
> amd64_prologue_line_bug initialization.
> * symtab.h (struct symtab): Remove field amd64_prologue_line_bug.
> * utils.c (producer_is_gcc_ge_4): Moved here from dwarf2read.c.
Checked in:
http://sourceware.org/ml/gdb-cvs/2011-09/msg00056.html
Thanks,
Jan
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [patch] workaround gcc46: prologue skip skips too far (PR 12435) #2
2011-09-09 19:52 ` Jan Kratochvil
@ 2011-09-12 17:23 ` Doug Evans
0 siblings, 0 replies; 8+ messages in thread
From: Doug Evans @ 2011-09-12 17:23 UTC (permalink / raw)
To: Jan Kratochvil; +Cc: gdb-patches
On Fri, Sep 9, 2011 at 12:41 PM, Jan Kratochvil
<jan.kratochvil@redhat.com> wrote:
> On Thu, 08 Sep 2011 20:22:27 +0200, Jan Kratochvil wrote:
>> gdb/
>> 2011-09-08 Jan Kratochvil <jan.kratochvil@redhat.com>
>>
>> Code cleanup.
>> * amd64-tdep.c (amd64_skip_prologue): Move the XMM code to ...
>> (amd64_skip_xmm_prologue): ... this new function. Describe its
>> parameters. No longer use amd64_prologue_line_bug.
>> * defs.h (producer_is_gcc_ge_4): New declaration.
>> * dwarf2read.c (producer_is_gcc_ge_4): Move to utils.c.
>> (process_full_comp_unit): Update its caller. Remove
>> amd64_prologue_line_bug initialization.
>> * symtab.h (struct symtab): Remove field amd64_prologue_line_bug.
>> * utils.c (producer_is_gcc_ge_4): Moved here from dwarf2read.c.
>
> Checked in:
> http://sourceware.org/ml/gdb-cvs/2011-09/msg00056.html
>
>
> Thanks,
> Jan
>
"works for me". Thanks.
^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2011-09-12 16:44 UTC | newest]
Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2011-07-22 23:04 [patch] workaround gcc46: prologue skip skips too far (PR 12435) #2 Jan Kratochvil
2011-07-24 18:16 ` Mark Kettenis
2011-07-24 20:00 ` Jan Kratochvil
2011-09-08 15:54 ` Jan Kratochvil
2011-09-08 17:31 ` Doug Evans
2011-09-08 19:13 ` Jan Kratochvil
2011-09-09 19:52 ` Jan Kratochvil
2011-09-12 17:23 ` Doug Evans
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox