This is the mail archive of the mailing list for the Cygwin project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [BUG] Intermittant error referencing memory @0x00000010

David A. Cobb writes:
> On 09/07/2002, David A. Cobb wrote:
> > Trying to do a bootstrap build of gcc-3.2-1, first cat.exe signalled
> > 'cat.exe the instruction at 0x077f8ed61 referenced memory at
> > 0x0000010. The memory could not be "written"'
> > followed by "Application failed to initialize (0xC0000142)"

0xC0000142 looks like a Windows exception code.  Keep in mind on Win9x
you'll get the infamous page fault error and no Application failed to
initialize msg.

> > Retry, got the same error but in "sh.exe instruction at 0x77f8ed61" --
> > HEY! that's strange - the same instruction address.
> > must be in a dll, yes?

Correct.  Apps don't have such high addresses, for future reference--they're
normally based at 0x400000 (so they can run on Windows 9x).

objdump on cygwin1.dll shows me one DLL dependency: kernel32.dll.  objdump
on kernel32 shows base is 0x77e80000.  But objdump on kernel32.dll also
shows it depends upon ntdll.dll.  objdump on ntdll.dll shows base is
0x77f80000, and has no DLL dependencies.  The code triggering the crash is
inside ntdll.dll.

Since kernel32.dll and ntdll.dll are *likely* OK (you never know M$, though,
but let's assume their code is kosher anyhow), AND considering that 3
SEPARATE .exe's caused the SAME problem, it looks like the only suspect left
is cygwin1.dll.

Are you using the latest Cygwin DLL version?  You can find out just what
that version number is from the output of "uname -a"; it's right after your

Eric R. Krause

Unsubscribe info:
Bug reporting:

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]