On 6/7/06, Christopher Smith <[EMAIL PROTECTED]> wrote:
Rick Funderburg wrote:
> On Jun 7, 2006, at 18:25 , Tracy R Reed wrote:
>
>> Does qmail or any other "bug free codes" written by arrogant
>> programmers implement a binary search or merge sort? Any programs
>> somehow use a sort in some way that might affect security? I wonder
>> if any fun can be had with this... Of course you need to introduce a
>> billion elements to be sorted so it might be a bit hard to exploit
>> but you never know!
>
> I can't say that I am often allocating 4 gigabytes of memory for a
> single variable, but I certainly wish I had enough memory to make it a
> non-event.
Actually, these kind of issues have been cropping up a fair bit in the
C/C++ world. They've been cleaning out bugs like this from the GNU C
core libraries for a number of years now. It's kind of disappointing
that Java's core libraries have bugs like this in them, and more
importantly that they are a "new" thing to the folks who wrote them.
--Chris
P.S.: Now that we've got 64-bit computing, 4GB is an increasingly
realistic address space for a process. For example, on the project I'm
working with at Yahoo, we're running 32-bit apps on a 64-bit kernel, and
we end up tweaking kernel parameters to get as close to using the entire
4GB address space as possible.
Now that we have 64-bit computing, we can run 64-bit apps, and integer
overflow happens at 2^63 not 2^31. So the safe range has increased by
a factor of 2^32, the well-known 4G number.
carl
--
carl lowenstein marine physical lab u.c. san diego
[EMAIL PROTECTED]
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg
