|
From: Dan K. <da...@ke...> - 2009-10-11 17:34:32
|
This weekend, I started valgrinding wine apps again,
and noticed strange new errors with 3.5.0 that I hadn't
seen before; it seems that valgrind isn't dealing with
mmap the same way it used to. I narrowed it down
to between 2009-05-27 and 2009-05-28, when it seems
a lot of darwin code landed.
For me, this causes (e.g.) the following wine test to break:
valgrind --trace-children=yes --track-origins=yes
--workaround-gcc296-bugs=yes ~/wine-git/wine advapi32_test.exe.so
cred.c
With the newer version of valgrind, this and probably all wine apps
complain very early with
Invalid write of size 4
at 0xCC56EDA: HEAP_CreateSubHeap (heap.c:815)
by 0xCC586D5: RtlCreateHeap (heap.c:1266)
by 0xCC8FBA1: thread_init (thread.c:328)
by 0xCC6192C: __wine_process_init (loader.c:2695)
by 0xC53091A: wine_init (loader.c:711)
by 0x7BF01274: main (main.c:218)
Address 0x7f00000c is not stack'd, malloc'd or (recently) free'd
Has anybody else noticed something funny about that time range?
|