Blame SOURCES/gdb-rhbz1909902-frame_id_p-assert-2.patch

405ea9
From FEDORA_PATCHES Mon Sep 17 00:00:00 2001
405ea9
From: Pedro Alves <pedro@palves.net>
405ea9
Date: Sat, 31 Oct 2020 00:27:18 +0000
405ea9
Subject: gdb-rhbz1909902-frame_id_p-assert-2.patch
405ea9
405ea9
;; Backport patch #2 which fixes a frame_id_p assertion failure (RH BZ 1909902).
405ea9
405ea9
Fix frame cycle detection
405ea9
405ea9
The recent commit to make scoped_restore_current_thread's cdtors
405ea9
exception free regressed gdb.base/eh_return.exp:
405ea9
405ea9
  Breakpoint 1, 0x00000000004012bb in eh2 (gdb/frame.c:641: internal-error: frame_id get_frame_id(frame_info*): Assertion `stashed' failed.
405ea9
  A problem internal to GDB has been detected,
405ea9
  further debugging may prove unreliable.
405ea9
  Quit this debugging session? (y or n) FAIL: gdb.base/eh_return.exp: hit breakpoint (GDB internal error)
405ea9
405ea9
That testcase uses __builtin_eh_return and, before the regression, the
405ea9
backtrace at eh2 looked like this:
405ea9
405ea9
 (gdb) bt
405ea9
 #0  0x00000000004006eb in eh2 (p=0x4006ec <continuation>) at src/gdb/testsuite/gdb.base/eh_return.c:54
405ea9
 Backtrace stopped: previous frame identical to this frame (corrupt stack?)
405ea9
405ea9
That "previous frame identical to this frame" is caught by the cycle
405ea9
detection based on frame id.
405ea9
405ea9
The assertion failing is this one:
405ea9
405ea9
 638           /* Since this is the first frame in the chain, this should
405ea9
 639              always succeed.  */
405ea9
 640           bool stashed = frame_stash_add (fi);
405ea9
 641           gdb_assert (stashed);
405ea9
405ea9
originally added by
405ea9
405ea9
  commit f245535cf583ae4ca13b10d47b3c7d3334593ece
405ea9
  Author:     Pedro Alves <palves@redhat.com>
405ea9
  AuthorDate: Mon Sep 5 18:41:38 2016 +0100
405ea9
405ea9
      Fix PR19927: Avoid unwinder recursion if sniffer uses calls parse_and_eval
405ea9
405ea9
The assertion is failing because frame #1's frame id was stashed
405ea9
before the id of frame #0 is stashed.  The frame id of frame #1 was
405ea9
stashed here:
405ea9
405ea9
 (top-gdb) bt
405ea9
 #0  frame_stash_add (frame=0x1e24c90) at src/gdb/frame.c:276
405ea9
 #1  0x0000000000669c1b in get_prev_frame_if_no_cycle (this_frame=0x19f8370) at src/gdb/frame.c:2120
405ea9
 #2  0x000000000066a339 in get_prev_frame_always_1 (this_frame=0x19f8370) at src/gdb/frame.c:2303
405ea9
 #3  0x000000000066a360 in get_prev_frame_always (this_frame=0x19f8370) at src/gdb/frame.c:2319
405ea9
 #4  0x000000000066b56c in get_frame_unwind_stop_reason (frame=0x19f8370) at src/gdb/frame.c:3028
405ea9
 #5  0x000000000059f929 in dwarf2_frame_cfa (this_frame=0x19f8370) at src/gdb/dwarf2/frame.c:1462
405ea9
 #6  0x00000000005ce434 in dwarf_evaluate_loc_desc::get_frame_cfa (this=0x7fffffffc070) at src/gdb/dwarf2/loc.c:666
405ea9
 #7  0x00000000005989a9 in dwarf_expr_context::execute_stack_op (this=0x7fffffffc070, op_ptr=0x1b2a053 "\364\003", op_end=0x1b2a053 "\364\003") at src/gdb/dwarf2/expr.c:1161
405ea9
 #8  0x0000000000596af6 in dwarf_expr_context::eval (this=0x7fffffffc070, addr=0x1b2a052 "\234\364\003", len=1) at src/gdb/dwarf2/expr.c:303
405ea9
 #9  0x0000000000597b4e in dwarf_expr_context::execute_stack_op (this=0x7fffffffc070, op_ptr=0x1b2a063 "", op_end=0x1b2a063 "") at src/gdb/dwarf2/expr.c:865
405ea9
 #10 0x0000000000596af6 in dwarf_expr_context::eval (this=0x7fffffffc070, addr=0x1b2a061 "\221X", len=2) at src/gdb/dwarf2/expr.c:303
405ea9
 #11 0x00000000005c8b5a in dwarf2_evaluate_loc_desc_full (type=0x1b564d0, frame=0x19f8370, data=0x1b2a061 "\221X", size=2, per_cu=0x1b28760, per_objfile=0x1a84930, subobj_type=0x1b564d0, subobj_byte_offset=0) at src/gdb/dwarf2/loc.c:2260
405ea9
 #12 0x00000000005c9243 in dwarf2_evaluate_loc_desc (type=0x1b564d0, frame=0x19f8370, data=0x1b2a061 "\221X", size=2, per_cu=0x1b28760, per_objfile=0x1a84930) at src/gdb/dwarf2/loc.c:2444
405ea9
 #13 0x00000000005cb769 in locexpr_read_variable (symbol=0x1b59840, frame=0x19f8370) at src/gdb/dwarf2/loc.c:3687
405ea9
 #14 0x0000000000663137 in language_defn::read_var_value (this=0x122ea60 <c_language_defn>, var=0x1b59840, var_block=0x0, frame=0x19f8370) at src/gdb/findvar.c:618
405ea9
 #15 0x0000000000663c3b in read_var_value (var=0x1b59840, var_block=0x0, frame=0x19f8370) at src/gdb/findvar.c:822
405ea9
 #16 0x00000000008c7d9f in read_frame_arg (fp_opts=..., sym=0x1b59840, frame=0x19f8370, argp=0x7fffffffc470, entryargp=0x7fffffffc490) at src/gdb/stack.c:542
405ea9
 #17 0x00000000008c89cd in print_frame_args (fp_opts=..., func=0x1b597c0, frame=0x19f8370, num=-1, stream=0x1aba860) at src/gdb/stack.c:890
405ea9
 #18 0x00000000008c9bf8 in print_frame (fp_opts=..., frame=0x19f8370, print_level=0, print_what=SRC_AND_LOC, print_args=1, sal=...) at src/gdb/stack.c:1394
405ea9
 #19 0x00000000008c92b9 in print_frame_info (fp_opts=..., frame=0x19f8370, print_level=0, print_what=SRC_AND_LOC, print_args=1, set_current_sal=1) at src/gdb/stack.c:1119
405ea9
 #20 0x00000000008c75f0 in print_stack_frame (frame=0x19f8370, print_level=0, print_what=SRC_AND_LOC, set_current_sal=1) at src/gdb/stack.c:366
405ea9
 #21 0x000000000070250b in print_stop_location (ws=0x7fffffffc9e0) at src/gdb/infrun.c:8110
405ea9
 #22 0x0000000000702569 in print_stop_event (uiout=0x1a8b9e0, displays=true) at src/gdb/infrun.c:8126
405ea9
 #23 0x000000000096d04b in tui_on_normal_stop (bs=0x1bcd1c0, print_frame=1) at src/gdb/tui/tui-interp.c:98
405ea9
 ...
405ea9
405ea9
Before the commit to make scoped_restore_current_thread's cdtors
405ea9
exception free, scoped_restore_current_thread's dtor would call
405ea9
get_frame_id on the selected frame, and we use
405ea9
scoped_restore_current_thread pervasively.  That had the side effect
405ea9
of stashing the frame id of frame #0 before reaching the path shown in
405ea9
the backtrace.  I.e., the frame id of frame #0 happened to be stashed
405ea9
before the frame id of frame #1.  But that was by chance, not by
405ea9
design.
405ea9
405ea9
This commit:
405ea9
405ea9
  commit 256ae5dbc73d1348850f86ee77a0dc3b04bc7cc0
405ea9
  Author:     Kevin Buettner <kevinb@redhat.com>
405ea9
  AuthorDate: Mon Oct 31 12:47:42 2016 -0700
405ea9
405ea9
      Stash frame id of current frame before stashing frame id for previous frame
405ea9
405ea9
Fixed a similar problem, by making sure get_prev_frame computes the
405ea9
frame id of the current frame before unwinding the previous frame, so
405ea9
that the cycle detection works properly.  That fix misses the scenario
405ea9
we're now running against, because if you notice, the backtrace above
405ea9
shows that frame #4 calls get_prev_frame_always, not get_prev_frame.
405ea9
I.e., nothing is calling get_frame_id on the current frame.
405ea9
405ea9
The fix here is to move Kevin's fix down from get_prev_frame to
405ea9
get_prev_frame_always.  Or actually, a bit further down to
405ea9
get_prev_frame_always_1 -- note that inline_frame_this_id calls
405ea9
get_prev_frame_always, so we need to be careful to avoid recursion in
405ea9
that scenario.
405ea9
405ea9
gdb/ChangeLog:
405ea9
405ea9
	* frame.c (get_prev_frame): Move get_frame_id call from here ...
405ea9
	(get_prev_frame_always_1): ... to here.
405ea9
	* inline-frame.c (inline_frame_this_id): Mention
405ea9
	get_prev_frame_always_1 in comment.
405ea9
405ea9
Change-Id: Id960c98ab2d072c48a436c3eb160cc4b2a5cfd1d
405ea9
405ea9
diff --git a/gdb/frame.c b/gdb/frame.c
405ea9
--- a/gdb/frame.c
405ea9
+++ b/gdb/frame.c
405ea9
@@ -2133,6 +2133,23 @@ get_prev_frame_always_1 (struct frame_info *this_frame)
405ea9
   if (get_frame_type (this_frame) == INLINE_FRAME)
405ea9
     return get_prev_frame_if_no_cycle (this_frame);
405ea9
 
405ea9
+  /* If this_frame is the current frame, then compute and stash its
405ea9
+     frame id prior to fetching and computing the frame id of the
405ea9
+     previous frame.  Otherwise, the cycle detection code in
405ea9
+     get_prev_frame_if_no_cycle() will not work correctly.  When
405ea9
+     get_frame_id() is called later on, an assertion error will be
405ea9
+     triggered in the event of a cycle between the current frame and
405ea9
+     its previous frame.
405ea9
+
405ea9
+     Note we do this after the INLINE_FRAME check above.  That is
405ea9
+     because the inline frame's frame id computation needs to fetch
405ea9
+     the frame id of its previous real stack frame.  I.e., we need to
405ea9
+     avoid recursion in that case.  This is OK since we're sure the
405ea9
+     inline frame won't create a cycle with the real stack frame.  See
405ea9
+     inline_frame_this_id.  */
405ea9
+  if (this_frame->level == 0)
405ea9
+    get_frame_id (this_frame);
405ea9
+
405ea9
   /* Check that this frame is unwindable.  If it isn't, don't try to
405ea9
      unwind to the prev frame.  */
405ea9
   this_frame->stop_reason
405ea9
@@ -2410,16 +2427,6 @@ get_prev_frame (struct frame_info *this_frame)
405ea9
      something should be calling get_selected_frame() or
405ea9
      get_current_frame().  */
405ea9
   gdb_assert (this_frame != NULL);
405ea9
-  
405ea9
-  /* If this_frame is the current frame, then compute and stash
405ea9
-     its frame id prior to fetching and computing the frame id of the
405ea9
-     previous frame.  Otherwise, the cycle detection code in
405ea9
-     get_prev_frame_if_no_cycle() will not work correctly.  When
405ea9
-     get_frame_id() is called later on, an assertion error will
405ea9
-     be triggered in the event of a cycle between the current
405ea9
-     frame and its previous frame.  */
405ea9
-  if (this_frame->level == 0)
405ea9
-    get_frame_id (this_frame);
405ea9
 
405ea9
   frame_pc_p = get_frame_pc_if_available (this_frame, &frame_pc);
405ea9
 
405ea9
diff --git a/gdb/inline-frame.c b/gdb/inline-frame.c
405ea9
--- a/gdb/inline-frame.c
405ea9
+++ b/gdb/inline-frame.c
405ea9
@@ -161,7 +161,8 @@ inline_frame_this_id (struct frame_info *this_frame,
405ea9
      real frame's this_id method.  So we must call
405ea9
      get_prev_frame_always.  Because we are inlined into some
405ea9
      function, there must be previous frames, so this is safe - as
405ea9
-     long as we're careful not to create any cycles.  */
405ea9
+     long as we're careful not to create any cycles.  See related
405ea9
+     comments in get_prev_frame_always_1.  */
405ea9
   *this_id = get_frame_id (get_prev_frame_always (this_frame));
405ea9
 
405ea9
   /* We need a valid frame ID, so we need to be based on a valid