Bug 509587

# cat  memcheck/tests/badfree3.vgtest.log
Wed Sep 17 07:30:16 2025 executing  VALGRIND_LIB=/usr/libexec/valgrind VALGRIND_LIB_INNER=/usr/libexec/valgrind /usr/bin/valgrind --command-line-only=yes --memcheck:leak-check=no --tool=memcheck   -q ./badfree  > badfree3.stdout.out 2> badfree3.stderr.out
Wed Sep 17 07:30:22 2025 rc 0
Wed Sep 17 07:30:22 2025 stdout:
Wed Sep 17 07:30:22 2025 stderr:
==8923== Invalid free() / delete / delete[] / realloc()
==8923==    at 0x4848A23: free (vg_replace_malloc.c:989)
==8923==    by 0x401157: main (badfree.c:12)
==8923==  Address 0x87654321 is not stack'd, malloc'd or (recently) free'd
==8923== 
==8923== Invalid free() / delete / delete[] / realloc()
==8923==    at 0x4848A23: free (vg_replace_malloc.c:989)
==8923==    by 0x40116B: main (badfree.c:15)
==8923==  Address 0x1fff00098c is on thread 1's stack
==8923==  in frame #1, created by main (badfree.c:7)
==8923== 
Wed Sep 17 07:30:22 2025 executing diff -u /dev/null badfree3.stdout.out > badfree3.stdout.diff
Wed Sep 17 07:30:22 2025 rc 0
Wed Sep 17 07:30:23 2025 executing ./filter_stderr badfree.c < badfree3.stderr.out > vg_regtest.tmp.743
Wed Sep 17 07:30:23 2025 rc 0
Wed Sep 17 07:30:23 2025 executing diff -u badfree3.stderr.exp badfree3.stderr.out > badfree3.stderr.diff
Wed Sep 17 07:30:23 2025 rc 256
--- badfree3.stderr.exp	2011-04-05 23:00:00.000000000 +0000
+++ badfree3.stderr.out	2025-09-17 07:30:23.637969970 +0000
@@ -1,11 +1,11 @@
 Invalid free() / delete / delete[] / realloc()
    at 0x........: free (vg_replace_malloc.c:...)
-   by 0x........: main (tests/badfree.c:12)
+   by 0x........: main (badfree.c:12)
  Address 0x........ is not stack'd, malloc'd or (recently) free'd
 
 Invalid free() / delete / delete[] / realloc()
    at 0x........: free (vg_replace_malloc.c:...)
-   by 0x........: main (tests/badfree.c:15)
+   by 0x........: main (badfree.c:15)
  Address 0x........ is on thread 1's stack
  in frame #1, created by main (badfree.c:7)

That is odd. So the difference is the reported backtrace doesn't contain "tests/badfree.c:15" but "badfree.c:15" (no directory/, just the file:line).

badfree3 is run with --fullpath-after=memcheck/ --fullpath-after=coregrind/ so one would expect the tests/ path (after the memcheck/).

Which compiler is this?

Could you maybe show: [eu-]readelf --debug-dump=line memcheck/tests/badfree

For me that gives:

 The Directory Table (offset 0x22, lines 2, columns 1):
  Entry	Name
  0	(indirect line string, offset: 0): /home/mark/src/valgrind/memcheck/tests
  1	(indirect line string, offset: 0x31): /usr/include

 The File Name Table (offset 0x30, lines 3, columns 2):
  Entry	Dir	Name
  0	0	(indirect line string, offset: 0x27): badfree.c
  1	0	(indirect line string, offset: 0x27): badfree.c
  2	1	(indirect line string, offset: 0x3e): stdlib.h

We use gcc 15.2.
$ readelf --debug-dump=line memcheck/tests/badfree
[snip]
The Directory Table (offset 0x237, lines 2, columns 1):
  Entry	Name
  0	(indirect line string, offset: 0): /usr/src/debug/glibc/2.42+git/csu
  1	(indirect line string, offset: 0x22): ../sysdeps/x86_64

 The File Name Table (offset 0x245, lines 2, columns 2):
  Entry	Dir	Name
  0	1	(indirect line string, offset: 0x11f): crtn.S
  1	1	(indirect line string, offset: 0x11f): crtn.S

 Line Number Statements:
  [0x0000024f]  Extended opcode 2: set Address to 0x401016
  [0x0000025a]  Advance Line by 39 to 40
  [0x0000025c]  Copy
  [0x0000025d]  Special opcode 62: advance Address by 4 to 0x40101a and Line by 1 to 41
  [0x0000025e]  Advance PC by 1 to 0x40101b
  [0x00000260]  Extended opcode 1: End of Sequence

  [0x00000263]  Extended opcode 2: set Address to 0x40118c
  [0x0000026e]  Advance Line by 43 to 44
  [0x00000270]  Copy
  [0x00000271]  Special opcode 62: advance Address by 4 to 0x401190 and Line by 1 to 45
  [0x00000272]  Advance PC by 1 to 0x401191
  [0x00000274]  Extended opcode 1: End of Sequence

Using GCC 15.2 on Rocky 8.10 I get output almost Identical to Mark's - only the path to my build directory is different.