[Bug 20315] New: PEB not fully initialized, causes valgrind warnings during thread initialization?

http://bugs.winehq.org/show_bug.cgi?id=20315

Summary: PEB not fully initialized, causes valgrind warnings
                    during thread initialization?
           Product: Wine
           Version: 1.1.31
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Keywords: download, patch, source
          Severity: normal
          Priority: P2
         Component: ntdll
        AssignedTo: [e-mail]
        ReportedBy: [e-mail]

Running wine's tests under valgrind yields warnings of the sort (after applying
the patch in bug 20303):

Syscall param socketcall.sendmsg(msg.msg_iov[i]) points to uninitialised
byte(s)
   at: sendmsg (socket.S:64)
   by: server_init_thread (server.c:1061)
   by: thread_init (thread.c:325)
   by: __wine_process_init (loader.c:2695)
   by: wine_init (loader.c:711)
   by: main (main.c:218)
 Address 0xfe924c48 is on thread 1's stack
 Uninitialised value was created by a client request
   at: inform_valgrind_of_new_virtual_memory_block (virtual.c:1725)
   by: NtAllocateVirtualMemory (virtual.c:1878)
   by: thread_init (thread.c:307)
   by: __wine_process_init (loader.c:2695)
   by: wine_init (loader.c:711)
   by: main (main.c:218)

and

Conditional jump or move depends on uninitialised value(s)
   at: get_modref (loader.c:293)
   by: load_dll (loader.c:1944)
   by: LdrLoadDll (loader.c:2025)
   by: load_library (module.c:890)
   by: LoadLibraryExW (module.c:947)
   by: __wine_kernel_init (process.c:1087)
   by: wine_init (loader.c:711)
   by: main (main.c:218)
 Uninitialised value was created by a client request
   at: inform_valgrind_of_new_virtual_memory_block (virtual.c:1725)
   by: NtAllocateVirtualMemory (virtual.c:1878)
   by: thread_init (thread.c:276)
   by: __wine_process_init (loader.c:2695)
   by: wine_init (loader.c:711)
   by: main (main.c:218)

I didn't look too far, but it seems that the PEB isn't fully
initialized and/or has holes that might need a valgrind annotation.
Memsetting them to zero right after allocation makes the 
warnings go away.  (The right fix would be to figure out which
fields still need initialization.)

[Bug 20315] PEB and TEB not fully initialized, causes valgrind warnings during thread initialization?

http://bugs.winehq.org/show_bug.cgi?id=20315

Dan Kegel  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|PEB not fully initialized,  |PEB and TEB not fully
                   |causes valgrind warnings    |initialized, causes
                   |during thread               |valgrind warnings during
                   |initialization?             |thread initialization?

--- Comment #1 from Dan Kegel   2009-10-10 21:38:30 ---
(meant to say PEB and TEB.  The patch memsets them both.)

[Bug 20315] PEB and TEB not fully initialized, causes valgrind warnings during thread initialization?

http://bugs.winehq.org/show_bug.cgi?id=20315

--- Comment #2 from Dan Kegel   2009-10-10 21:39:21 ---
Created an attachment (id=24021)
 --> (http://bugs.winehq.org/attachment.cgi?id=24021)
Quick patch to work around missing initialization

[Bug 20315] PEB and TEB not fully initialized, causes valgrind warnings during thread initialization?

http://bugs.winehq.org/show_bug.cgi?id=20315

Alexandre Julliard  changed:

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

--- Comment #3 from Alexandre Julliard   2009-10-11 03:55:05 ---
mmap initializes to zero, your annotations are broken.

[Bug 20315] PEB and TEB not fully initialized, causes valgrind warnings during thread initialization?

http://bugs.winehq.org/show_bug.cgi?id=20315

Dmitry Timoshkov  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |CLOSED

--- Comment #4 from Dmitry Timoshkov   2009-10-11 07:49:37 ---
Closing invalid.