[PATCH, v2] Fix reverse stepping multiple contiguous PC ranges over the line table

[Luis Machado <luis.machado@arm.com>]

v2:
- Check if both the line and symtab match for a particular line table entry.

--

When running GDB's testsuite on aarch64-linux/Ubuntu 20.04 (also spotted on
the ppc backend), I noticed some failures in gdb.reverse/solib-precsave.exp
and gdb.reverse/solib-reverse.exp.

The failure happens around the following code:

38  b[1] = shr2(17);		/* middle part two */
40  b[0] = 6;   b[1] = 9;	/* generic statement, end part two */
42  shr1 ("message 1\n");	/* shr1 one */

Normal execution:

- step from line 38 will land on line 40.
- step from line 40 will land on line 42.

Reverse execution:

- step from line 42 will land on line 40.
- step from line 40 will land on line 40.
- step from line 40 will land on line 38.

The problem here is that line 40 contains two contiguous but distinct
PC ranges in the line table, like so:

Line 40 - [0x7ec ~ 0x7f4]
Line 40 - [0x7f4 ~ 0x7fc]

The two distinct ranges are generated because GCC started outputting source
column information, which GDB doesn't take into account at the moment.

When stepping forward from line 40, we skip both of these ranges and land on
line 42. When stepping backward from line 42, we stop at the start PC of the
second (or first, going backwards) range of line 40.

This happens because we have this check in infrun.c:process_event_stop_test:

      /* When stepping backward, stop at beginning of line range
	 (unless it's the function entry point, in which case
	 keep going back to the call point).  */
      CORE_ADDR stop_pc = ecs->event_thread->stop_pc ();
      if (stop_pc == ecs->event_thread->control.step_range_start
	  && stop_pc != ecs->stop_func_start
	  && execution_direction == EXEC_REVERSE)
	end_stepping_range (ecs);
      else
	keep_going (ecs);

Since we've reached ecs->event_thread->control.step_range_start, we stop
stepping backwards.

The right thing to do is to look for adjacent PC ranges for the same line,
until we notice a line change. Then we take that as the start PC of the
range.

Another solution I thought about is to merge the contiguous ranges when
we are reading the line tables. Though I'm not sure if we really want to process
that data as opposed to keeping it as the compiler created, and then working
around that.

In any case, the following patch addresses this problem.

I'm not particularly happy with how we go back in the ranges (using "pc - 1").
Feedback would be welcome.

Validated on aarch64-linux/Ubuntu 20.04/18.04. Carl Love has verified that it
does fix a similar issue on ppc.

Ubuntu 18.04 doesn't actually run into these failures because the compiler
doesn't generate distinct PC ranges for the same line.

I see similar failures on x86_64 in the gdb.reverse tests
(gdb.reverse/step-reverse.exp and gdb.reverse/step-reverse.exp). Those are
also fixed by this patch, although Carl's testcase doesn't fail for x86_64.

There seems to be a corner case where a line table has only one instruction,
and while stepping that doesn't take the same path through infrun. I think it
needs some more investigation. Therefore this is a tentative patch for now.

Co-authored-by: Carl Love <cel@us.ibm.com>
---
 gdb/infrun.c                                  |  51 +++++++-
 gdb/symtab.c                                  |  49 ++++++++
 gdb/symtab.h                                  |  16 +++
 gdb/testsuite/gdb.reverse/map-to-same-line.c  |  30 +++++
 .../gdb.reverse/map-to-same-line.exp          | 114 ++++++++++++++++++
 5 files changed, 258 insertions(+), 2 deletions(-)
 create mode 100644 gdb/testsuite/gdb.reverse/map-to-same-line.c
 create mode 100644 gdb/testsuite/gdb.reverse/map-to-same-line.exp

diff --git a/gdb/infrun.c b/gdb/infrun.c
index dbd82775595..cec55b0a5d7 100644
--- a/gdb/infrun.c
+++ b/gdb/infrun.c
@@ -6716,6 +6716,11 @@ process_event_stop_test (struct execution_control_state *ecs)
      through a function epilogue and therefore must detect when
      the current-frame changes in the middle of a line.  */
 
+      infrun_debug_printf
+	("STEPPING RANGE [%s-%s]",
+	 paddress (gdbarch, ecs->event_thread->control.step_range_start),
+	 paddress (gdbarch, ecs->event_thread->control.step_range_end));
+
   if (pc_in_thread_step_range (ecs->event_thread->stop_pc (),
 			       ecs->event_thread)
       && (execution_direction != EXEC_REVERSE
@@ -6723,7 +6728,7 @@ process_event_stop_test (struct execution_control_state *ecs)
 			  ecs->event_thread->control.step_frame_id)))
     {
       infrun_debug_printf
-	("stepping inside range [%s-%s]",
+	("STILL INSIDE range [%s-%s]",
 	 paddress (gdbarch, ecs->event_thread->control.step_range_start),
 	 paddress (gdbarch, ecs->event_thread->control.step_range_end));
 
@@ -6732,11 +6737,31 @@ process_event_stop_test (struct execution_control_state *ecs)
 	 have software watchpoints).  */
       ecs->event_thread->control.may_range_step = 1;
 
+      /* When we are stepping inside a particular line range, in reverse,
+	 and we are sitting at the first address of that range, we need to
+	 check if this address also shows up in another line range as the
+	 end address.
+
+	 If so, we need to check what line such a step range points to.
+	 If it points to the same line as the current step range, that
+	 means we need to keep going in order to reach the first address
+	 of the line range.  We repeat this until we eventually get to the
+	 first address of a particular line we're stepping through.  */
+      CORE_ADDR range_start = ecs->event_thread->control.step_range_start;
+      if (execution_direction == EXEC_REVERSE)
+	{
+	  gdb::optional<CORE_ADDR> real_range_start
+	    = find_line_range_start (ecs->event_thread->stop_pc ());
+
+	  if (real_range_start.has_value ())
+	    range_start = *real_range_start;
+	}
+
       /* When stepping backward, stop at beginning of line range
 	 (unless it's the function entry point, in which case
 	 keep going back to the call point).  */
       CORE_ADDR stop_pc = ecs->event_thread->stop_pc ();
-      if (stop_pc == ecs->event_thread->control.step_range_start
+      if (stop_pc == range_start
 	  && stop_pc != ecs->stop_func_start
 	  && execution_direction == EXEC_REVERSE)
 	end_stepping_range (ecs);
@@ -7237,6 +7262,28 @@ process_event_stop_test (struct execution_control_state *ecs)
 	}
     }
 
+    CORE_ADDR range_start = ecs->event_thread->control.step_range_start;
+    infrun_debug_printf ("RANGE START is at %s", paddress (gdbarch, range_start));
+    if (execution_direction == EXEC_REVERSE)
+      {
+	gdb::optional<CORE_ADDR> real_range_start
+	  = find_line_range_start (ecs->event_thread->stop_pc ());
+
+	if (real_range_start.has_value ())
+	  {
+	    range_start = *real_range_start;
+	    infrun_debug_printf ("NEW RANGE START is at %s", paddress (gdbarch, range_start));
+	  }
+
+	/* When stepping backward, stop at beginning of line range
+	   (unless it's the function entry point, in which case
+	   keep going back to the call point).  */
+	CORE_ADDR stop_pc = ecs->event_thread->stop_pc ();
+	if (stop_pc == range_start
+	    && stop_pc != ecs->stop_func_start)
+	  end_stepping_range (ecs);
+      }
+
   /* We aren't done stepping.
 
      Optimize by setting the stepping range to the line.
diff --git a/gdb/symtab.c b/gdb/symtab.c
index 804b88284eb..99abf7bd2b7 100644
--- a/gdb/symtab.c
+++ b/gdb/symtab.c
@@ -3425,6 +3425,55 @@ find_pc_line (CORE_ADDR pc, int notcurrent)
   return sal;
 }
 
+/* Compare two symtab_and_line entries.  Return true if both have
+   the same line number and the same symtab pointer.  That means we
+   are dealing with two entries from the same line and from the same
+   source file.
+
+   Return false otherwise.  */
+
+static bool
+sal_line_symtab_matches_p (const symtab_and_line &sal1,
+			   const symtab_and_line &sal2)
+{
+  return (sal1.line == sal2.line && sal1.symtab == sal2.symtab);
+}
+
+/* See symtah.h.  */
+
+gdb::optional<CORE_ADDR>
+find_line_range_start (CORE_ADDR pc)
+{
+  struct symtab_and_line current_sal = find_pc_line (pc, 0);
+
+  if (current_sal.line == 0)
+    return {};
+
+  struct symtab_and_line prev_sal = find_pc_line (current_sal.pc - 1, 0);
+
+  /* If the previous entry is for a different line, that means we are already
+     at the entry with the start PC for this line.  */
+  if (!sal_line_symtab_matches_p (prev_sal, current_sal))
+    return current_sal.pc;
+
+  /* Otherwise, keep looking for entries for the same line but with
+     smaller PC's.  */
+  bool done = false;
+  CORE_ADDR prev_pc;
+  while (!done)
+    {
+      prev_pc = prev_sal.pc;
+
+      prev_sal = find_pc_line (prev_pc - 1, 0);
+
+      /* Did we notice a line change?  If so, we are done with the search.  */
+      if (!sal_line_symtab_matches_p (prev_sal, current_sal))
+	done = true;
+    }
+
+  return prev_pc;
+}
+
 /* See symtab.h.  */
 
 struct symtab *
diff --git a/gdb/symtab.h b/gdb/symtab.h
index d12eee6e9d8..4d893a8a3b8 100644
--- a/gdb/symtab.h
+++ b/gdb/symtab.h
@@ -2149,6 +2149,22 @@ extern struct symtab_and_line find_pc_line (CORE_ADDR, int);
 extern struct symtab_and_line find_pc_sect_line (CORE_ADDR,
 						 struct obj_section *, int);
 
+/* Given PC, and assuming it is part of a range of addresses that is part of a
+   line, go back through the linetable and find the starting PC of that
+   line.
+
+   For example, suppose we have 3 PC ranges for line X:
+
+   Line X - [0x0 - 0x8]
+   Line X - [0x8 - 0x10]
+   Line X - [0x10 - 0x18]
+
+   If we call the function with PC == 0x14, we want to return 0x0, as that is
+   the starting PC of line X, and the ranges are contiguous.
+*/
+
+extern gdb::optional<CORE_ADDR> find_line_range_start (CORE_ADDR pc);
+
 /* Wrapper around find_pc_line to just return the symtab.  */
 
 extern struct symtab *find_pc_line_symtab (CORE_ADDR);
diff --git a/gdb/testsuite/gdb.reverse/map-to-same-line.c b/gdb/testsuite/gdb.reverse/map-to-same-line.c
new file mode 100644
index 00000000000..d35838eccc9
--- /dev/null
+++ b/gdb/testsuite/gdb.reverse/map-to-same-line.c
@@ -0,0 +1,30 @@
+
+
+/* The purpose of this test is to create a DWARF line table that says there
+   are two assignment statements on the same line.  The expect test checks to
+   make sure gdb reverse steps over each statements individulally.  The test
+   makes sure a single reverse step doesn't step over both assignment
+   statements in the line. The expect file generates the DWARF assembly
+   statements will create a line table that with j = 3 and k = 4 listed as
+   being on the same source code line even though they actually are on
+   different lines in this source code below. Have to put them on separate
+   lines to be able to identify them and add them to the line table.  */
+int
+main (){     /* TAG: main prologue */
+  asm ("main_label: .globl main_label");
+  asm ("loop_start: .globl loop_start");
+  int i, j, k, l, m;
+
+  asm ("i_assignment: .globl i_assignment");
+  i = 1;     /* TAG: assignment i */
+  asm ("j_assignment: .globl j_assignment");
+  j = 3;     /* TAG: assignment j */
+  asm ("k_assignment: .globl k_assignment");
+  k = 4;  /* TAG: assignment k */
+  asm ("l_assignment: .globl l_assignment");
+  l = 10;     /* TAG: assignment l */
+  asm ("m_assignment: .globl m_assignment");
+  m = 11;     /* TAG: assignment m */
+  asm ("main_return: .globl main_return");
+  return 0;  /* TAG: end of main */
+}
diff --git a/gdb/testsuite/gdb.reverse/map-to-same-line.exp b/gdb/testsuite/gdb.reverse/map-to-same-line.exp
new file mode 100644
index 00000000000..da407b50e94
--- /dev/null
+++ b/gdb/testsuite/gdb.reverse/map-to-same-line.exp
@@ -0,0 +1,114 @@
+
+# The purpose of this test is to create a DWARF line table that says the
+# two assignment statements are on the same line.  The expect test checks
+# to make sure gdb reverse steps over each statement individually, i.e.
+# not over the line containing both assignments.  */
+
+load_lib dwarf.exp
+
+# This test can only be run on targets which support DWARF-2 and use gas.
+if {![dwarf2_support]} {
+    unsupported "dwarf2 support required for this test"
+    return 0
+}
+
+if [get_compiler_info] {
+    return -1
+}
+
+# The DWARF assembler requires the gcc compiler.
+if {!$gcc_compiled} {
+    unsupported "gcc is required for this test"
+    return 0
+}
+
+standard_testfile .c .S
+
+if { [prepare_for_testing "failed to prepare" ${testfile} ${srcfile}] } {
+    return -1
+}
+
+set asm_file [standard_output_file $srcfile2]
+Dwarf::assemble $asm_file {
+    global srcdir subdir srcfile
+    declare_labels integer_label L
+
+    # Find start address and length of program
+    lassign [function_range main [list ${srcdir}/${subdir}/$srcfile]] \
+        main_start main_len
+    set main_end "$main_start + $main_len"
+
+    cu {} {
+        compile_unit {
+            {language @DW_LANG_C}
+            {name map-to-same-line.c}
+            {stmt_list $L DW_FORM_sec_offset}
+            {low_pc 0 addr}
+        } {
+            subprogram {
+                {external 1 flag}
+                {name main}
+                {low_pc $main_start addr}
+                {high_pc $main_len DW_FORM_data4}
+            }
+        }
+    }
+
+    lines {version 2 default_is_stmt 1} L {
+        include_dir "${srcdir}/${subdir}"
+	file_name "$srcfile" 1
+
+	# Generate the line table program with the assignments to j and k
+	# are listed on the same source line in the table.
+	program {
+	    {DW_LNE_set_address $main_start}
+            {line [gdb_get_line_number "TAG: main prologue"]}
+	    {DW_LNS_copy}
+
+	    {DW_LNE_set_address i_assignment}
+	    {line [gdb_get_line_number "TAG: assignment i"]}
+	    {DW_LNS_copy}
+
+	    {DW_LNE_set_address j_assignment}
+	    {line [gdb_get_line_number "TAG: assignment j"]}
+	    {DW_LNS_copy}
+
+	    {DW_LNE_set_address k_assignment}
+	    {line [gdb_get_line_number "TAG: assignment j"]}
+	    {DW_LNS_copy}
+
+	    {DW_LNE_set_address l_assignment}
+	    {line [gdb_get_line_number "TAG: assignment l"]}
+	    {DW_LNS_copy}
+
+	    {DW_LNE_set_address m_assignment}
+	    {line [gdb_get_line_number "TAG: assignment m"]}
+	    {DW_LNS_copy}
+
+	    {DW_LNE_end_sequence}
+	}
+    }
+}
+
+if { [prepare_for_testing "failed to prepare" ${testfile} \
+        [list $srcfile $asm_file] {nodebug} ] } {
+    return -1
+}
+
+if ![runto_main] {
+    return -1
+}
+
+if [supports_process_record] {
+    # Activate process record/replay
+    gdb_test_no_output "record" "turn on process record"
+}
+
+set break_at_l [gdb_get_line_number "TAG: assignment l" ]
+
+gdb_test "tbreak $break_at_l" "Temporary breakpoint .*"  "breakpoint l = 10"
+
+gdb_test "continue" "Temporary breakpoint .*" "run to l = 10"
+# j = 3 and k = 4 are on the same line.  The reverse step stops at j = 3
+gdb_test "reverse-step" ".* j = 3;.*" "reverse-step to j = 3"
+gdb_test "reverse-step" ".* i = 1;.*" "reverse-step to i = 1"
-- 
2.25.1