public inbox for dwz@sourceware.org
 help / color / mirror / Atom feed
* [Bug default/25439] New: namespace DIE added to duplicate chain list (--odr-mode=basic)
@ 2020-01-01  0:00 vries at gcc dot gnu.org
  2020-01-01  0:00 ` [Bug default/25439] " vries at gcc dot gnu.org
                   ` (2 more replies)
  0 siblings, 3 replies; 4+ messages in thread
From: vries at gcc dot gnu.org @ 2020-01-01  0:00 UTC (permalink / raw)
  To: dwz

https://sourceware.org/bugzilla/show_bug.cgi?id=25439

            Bug ID: 25439
           Summary: namespace DIE added to duplicate chain list
                    (--odr-mode=basic)
           Product: dwz
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: default
          Assignee: nobody at sourceware dot org
          Reporter: vries at gcc dot gnu.org
                CC: dwz at sourceware dot org
  Target Milestone: ---

Starting commit 62e4949 "Fix reference from PU to CU (--odr-mode=basic)", we
run into an abort in build_abbrevs_for_die with:
...
$ dwz -Lnone -lnone -o 1 Checker.so.debug --odr-mode=basic
...

This is a consequence of calling partition_found_dups with a die where
die->named_namespace == 1 during mark_singletons, the function added in the
offending commit.

This happens because a die:
...
 (gdb) call dump_die (die)
 4db7a86 O ffe1edf 0  imported_module
...
has a reference to a namespace:
...
(gdb) call dump_die (reft)
 4c3e954 X 9a045343 0 llvm namespace
...

-- 
You are receiving this mail because:
You are on the CC list for the bug.

^ permalink raw reply	[flat|nested] 4+ messages in thread

* [Bug default/25439] namespace DIE added to duplicate chain list (--odr-mode=basic)
  2020-01-01  0:00 [Bug default/25439] New: namespace DIE added to duplicate chain list (--odr-mode=basic) vries at gcc dot gnu.org
@ 2020-01-01  0:00 ` vries at gcc dot gnu.org
  2020-01-01  0:00 ` vries at gcc dot gnu.org
  2021-02-23 13:11 ` vries at gcc dot gnu.org
  2 siblings, 0 replies; 4+ messages in thread
From: vries at gcc dot gnu.org @ 2020-01-01  0:00 UTC (permalink / raw)
  To: dwz

https://sourceware.org/bugzilla/show_bug.cgi?id=25439

--- Comment #2 from Tom de Vries <vries at gcc dot gnu.org> ---
Tentative patch:
...
index 08e3df4..e2a4ec4 100644
--- a/dwz.c
+++ b/dwz.c
@@ -3370,24 +3370,31 @@ checksum_ref_die (dw_cu_ref cu, dw_die_ref top_die,
dw_die_ref die,
        }
     }

-  if (!only_hash_name_p
-      && (top_die == NULL || top_die->die_ck_state != CK_BAD))
-    for (child = die->die_child; child; child = child->die_sib)
-      {
-       unsigned int r
-         = checksum_ref_die (cu,
-                             top_die ? top_die
-                             : child->die_named_namespace
-                             ? NULL : child, child,
-                             second_idx, second_hash);
-       if (top_die == NULL)
-         assert (r == 0 && obstack_object_size (&ob) == 0);
-
-       if (ret == 0 || (r && r < ret))
-         ret = r;
-       if (top_die && top_die->die_ck_state == CK_BAD)
-         break;
-      }
+  if (top_die == NULL || top_die->die_ck_state != CK_BAD)
+    {
+      hashval_t save_checksum;
+      if (top_die)
+       save_checksum = top_die->u.p1.die_ref_hash;
+      for (child = die->die_child; child; child = child->die_sib)
+       {
+         unsigned int r
+           = checksum_ref_die (cu,
+                               top_die ? top_die
+                               : child->die_named_namespace
+                               ? NULL : child, child,
+                               second_idx, second_hash);
+         if (top_die == NULL)
+           assert (r == 0 && obstack_object_size (&ob) == 0);
+
+         if (ret == 0 || (r && r < ret))
+           ret = r;
+         if (top_die && top_die->die_ck_state == CK_BAD)
+           break;
+       }
+      if (top_die
+         && top_die->die_ck_state != CK_BAD && only_hash_name_p)
+       top_die->u.p1.die_ref_hash = save_checksum;
+    }--

   if (top_die == die)
     {
...

However, that also drops effectiveness to less than 1%:
...
$ rm -f 1; ./dwz -Lnone -lnone -o 1 cc1 --no-odr
$ rm -f 2; ./dwz -Lnone -lnone -o 2 cc1 --odr-mode=basic
$ diff.sh cc1 1
.debug_info      red: 44.84%    111527248  61527733
.debug_abbrev    red: 40.28%    1722726  1028968
.debug_str       red: 0%        6609355  6609355
total            red: 42.30%    119859329 69166056
$ diff.sh cc1 2
.debug_info      red: 45.44%    111527248  60860046
.debug_abbrev    red: 42.87%    1722726  984340
.debug_str       red: 0%        6609355  6609355
total            red: 42.89%    119859329 68453741
...

-- 
You are receiving this mail because:
You are on the CC list for the bug.

^ permalink raw reply	[flat|nested] 4+ messages in thread

* [Bug default/25439] namespace DIE added to duplicate chain list (--odr-mode=basic)
  2020-01-01  0:00 [Bug default/25439] New: namespace DIE added to duplicate chain list (--odr-mode=basic) vries at gcc dot gnu.org
  2020-01-01  0:00 ` [Bug default/25439] " vries at gcc dot gnu.org
@ 2020-01-01  0:00 ` vries at gcc dot gnu.org
  2021-02-23 13:11 ` vries at gcc dot gnu.org
  2 siblings, 0 replies; 4+ messages in thread
From: vries at gcc dot gnu.org @ 2020-01-01  0:00 UTC (permalink / raw)
  To: dwz

https://sourceware.org/bugzilla/show_bug.cgi?id=25439

--- Comment #1 from Tom de Vries <vries at gcc dot gnu.org> ---
(In reply to Tom de Vries from comment #0)
> Starting commit 62e4949 "Fix reference from PU to CU (--odr-mode=basic)", we
> run into an abort in build_abbrevs_for_die with:
> ...
> $ dwz -Lnone -lnone -o 1 Checker.so.debug --odr-mode=basic
> ...
> 
> This is a consequence of calling partition_found_dups with a die where
> die->named_namespace == 1 during mark_singletons, the function added in the
> offending commit.
> 
> This happens because a die:
> ...
>  (gdb) call dump_die (die)
>  4db7a86 O ffe1edf 0  imported_module
> ...
> has a reference to a namespace:
> ...
> (gdb) call dump_die (reft)
>  4c3e954 X 9a045343 0 llvm namespace
> ...

And this again is a consequence of something we can detect earlier:
...
@@ -6343,6 +6350,7 @@ partition_cmp (const void *p, const void *q)
 static void
 partition_found_dups (dw_die_ref die, struct obstack *vec)
 {
+  assert (die->die_ck_state == CK_KNOWN);
   obstack_ptr_grow (vec, die);
   if (unlikely (verify_dups_p))
     verify_dups (die, true);
...

-- 
You are receiving this mail because:
You are on the CC list for the bug.

^ permalink raw reply	[flat|nested] 4+ messages in thread

* [Bug default/25439] namespace DIE added to duplicate chain list (--odr-mode=basic)
  2020-01-01  0:00 [Bug default/25439] New: namespace DIE added to duplicate chain list (--odr-mode=basic) vries at gcc dot gnu.org
  2020-01-01  0:00 ` [Bug default/25439] " vries at gcc dot gnu.org
  2020-01-01  0:00 ` vries at gcc dot gnu.org
@ 2021-02-23 13:11 ` vries at gcc dot gnu.org
  2 siblings, 0 replies; 4+ messages in thread
From: vries at gcc dot gnu.org @ 2021-02-23 13:11 UTC (permalink / raw)
  To: dwz

https://sourceware.org/bugzilla/show_bug.cgi?id=25439

Tom de Vries <vries at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |DUPLICATE
             Status|NEW                         |RESOLVED

--- Comment #3 from Tom de Vries <vries at gcc dot gnu.org> ---
Retested master, doesn't trigger anymore.  Bisect to fix for PR26252.

Marking it a duplicate.

*** This bug has been marked as a duplicate of bug 26252 ***

-- 
You are receiving this mail because:
You are on the CC list for the bug.

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2021-02-23 13:11 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-01-01  0:00 [Bug default/25439] New: namespace DIE added to duplicate chain list (--odr-mode=basic) vries at gcc dot gnu.org
2020-01-01  0:00 ` [Bug default/25439] " vries at gcc dot gnu.org
2020-01-01  0:00 ` vries at gcc dot gnu.org
2021-02-23 13:11 ` vries at gcc dot gnu.org

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).