Re: Invalid lock sequence
From: Ivan Brugiolo [MSFT] (ivanbrug_at_online.microsoft.com)
Date: 02/25/04
- Next message: David Crow [MCSD]: "Outputting text onto a bitmap"
- Previous message: Tim Wallace: "error creating obj using new"
- In reply to: Andrew: "Re: Invalid lock sequence"
- Messages sorted by: [ date ] [ thread ]
Date: Wed, 25 Feb 2004 10:23:39 -0800
If you are going after a reuse-after-delete,
you can try to use Full-PageHeap,
that will decommitt the block of memory,
so that any reuse-attempt will cause an AV.
if you ar also using one of cdb/ntsd/windbg as the debugger,
you can use
0:001>!heap -p -a <address_of_the_AV>
to get a stack trace of the code-path that free-ed the block of memory.
-- This posting is provided "AS IS" with no warranties, and confers no rights. Use of any included script samples are subject to the terms specified at http://www.microsoft.com/info/cpyright.htm "Andrew" <anonymous@discussions.microsoft.com> wrote in message news:D46D93CC-5E2E-47E9-B7CB-8DB8AD46B427@microsoft.com... > The code in question does use vtables, so it's possible that's what's happening. The call stack generally looked good, although there were a couple things that didn't quite match. I assumed this was due to the release build optimizations, but execution of arbitrary memory seems likely considering the results I'm getting. I will investigate further. > > Thanks! > > ----- Ivan Brugiolo [MSFT] wrote: ----- > > Are you sure you are not executing code out of arbitrary memory ? > Sometimes it happens when you perform a reuse-after-delete of an object with > a vtable. > The vtable is reused-memory, and it will point to arbitrary code. > Other cause would be a stack corruption that will make you execute arbitrary > code as well. > Single bit error in the memory page could be a cause as well. > > -- > This posting is provided "AS IS" with no warranties, and confers no rights. > Use of any included script samples are subject to the terms specified at > http://www.microsoft.com/info/cpyright.htm > > > "Andrew" <anonymous@discussions.microsoft.com> wrote in message > news:3CDCE1FC-FB6C-4E60-A7B9-C74D60C6754F@microsoft.com... > > I just encountered a bug in a release build of my code. According to the > exception, "An attempt was made to execute an invalid lock sequence." The > line of code that execution breaks on is: > >> lock fstp dword ptr [esp] > >> According to Intel's instruction set reference, such an operation will > generate an undefined opcode exception, because the FSTP instruction is not > in the list of instructions that LOCK can be prefixed to. So either the > compiler is generating invalid code, the assembly output is wrong, the Intel > documentation is incomplete, or I don't understand how the LOCK instruction > can be used. > >> Can anyone confirm whether or not this is a valid instruction? If this is > not a valid instruction, is anyone aware of this being a known issue with > the VC 7.1 compiler? > >> Thanks, > > Andrew > > >
- Next message: David Crow [MCSD]: "Outputting text onto a bitmap"
- Previous message: Tim Wallace: "error creating obj using new"
- In reply to: Andrew: "Re: Invalid lock sequence"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|
