https://bugs.winehq.org/show_bug.cgi?id=38673
Bug ID: 38673 Summary: valgrind shows uninitialized memory in dlls/kernel32/sync.c:ConnectNamedPipe() Product: Wine Version: 1.7.42 Hardware: x86 OS: Linux Status: NEW Keywords: download, source, testcase, valgrind Severity: normal Priority: P2 Component: kernel32 Assignee: wine-bugs@winehq.org Reporter: austinenglish@gmail.com Distribution: Fedora
../../../tools/runtest -q -P wine -T ../../.. -M hlink.dll -p hlink_test.exe.so hlink && touch hlink.ok ==14126== Thread 3: ==14126== Syscall param writev(vector[...]) points to uninitialised byte(s) ==14126== at 0x4F7A0F4B: ??? (in /usr/lib/libc-2.21.so) ==14126== by 0x7BC7A9C5: send_request (server.c:228) ==14126== by 0x7BC7AB52: wine_server_call (server.c:309) ==14126== by 0x7BC7B8CD: server_select (server.c:607) ==14126== by 0x7BC85426: NtWaitForMultipleObjects (sync.c:976) ==14126== by 0x7BC85471: NtWaitForSingleObject (sync.c:985) ==14126== by 0x7BC45DFF: server_ioctl_file (file.c:1539) ==14126== by 0x7BC46565: NtFsControlFile (file.c:1786) ==14126== by 0x7B87406B: ConnectNamedPipe (sync.c:1615) ==14126== by 0x49255A5: listen_thread (rpc_transport.c:132) ==14126== by 0x7BC81BB7: ??? (signal_i386.c:2682) ==14126== by 0x7BC81BFE: call_thread_func (signal_i386.c:2741) ==14126== by 0x7BC81B95: ??? (signal_i386.c:2682) ==14126== by 0x7BC88D78: start_thread (thread.c:443) ==14126== by 0x4F88F37F: start_thread (in /usr/lib/libpthread-2.21.so) ==14126== by 0x4F7ABD4D: clone (in /usr/lib/libc-2.21.so) ==14126== Address 0x4cee5c4 is on thread 1's stack ==14126== Uninitialised value was created by a stack allocation ==14126== at 0x7B873FA5: ConnectNamedPipe (sync.c:1601) ==14126==
also shows in ieframe/webbrowser, kernel32/pipe,
https://bugs.winehq.org/show_bug.cgi?id=38673
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Version|1.7.42 |1.7.44
https://bugs.winehq.org/show_bug.cgi?id=38673
Jacek Caban jacek@codeweavers.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |jacek@codeweavers.com
--- Comment #1 from Jacek Caban jacek@codeweavers.com --- It's probably fixed. I guess the bug was about accessing IO_STATUS_BLOCK, but this works differently in ntdll now. Also, listen_thread from this backtrace does not exist in rpcrt4 anymore.
https://bugs.winehq.org/show_bug.cgi?id=38673
François Gouget fgouget@codeweavers.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |fgouget@codeweavers.com Assignee|wine-bugs@winehq.org |austinenglish@gmail.com
--- Comment #2 from François Gouget fgouget@codeweavers.com --- Austin, does this still happen?