Re: [HACKERS] [PATCHES] fork/exec patch

> OK, you have a point. There is even one more thing that I thought
about in that services are
> restricted in message handling capabilities in NT unless you
explicitly disable that protection
> (basically a security fix for the scatter attack). So messages are
probably out, unless this
> restriction does not apply to 'child' processes which off the top of
my head I am not sure about.
> However scatter attack type problems do not really apply because the
postmaster would only listen
> for private messages (or public ones and implement the standard
behavior). IMO, the scatter
> problem was overhyped.

I haven't used message handlers in a long time, but can you use them at
all unless you are on the desktop? Meaning the Interact With Desktop
switch would be required, which in turn rules out any other user than
Local System. And that's the same as running as root on Unix, which is
something postgresql refuses.
Postgresql on windows shuold *definitly* support running with a low
privilege account.

> Also, storing the signal type in shared memory (as in the
> CONNX code) seems overkill because the signal ID (along with
> the pid) can be stored in the name of the event. Here is
> Microsoft's example (from the above
> link) on emulating SIGALRM with events:

The problem with this is that you eat up a lot of handles if you have a
lot of backends. When you haev a few, it's not an issue, but when you
start creating logs of backends, it might become noticeable.

> #define _WIN32_WINNT 0X0500
This means the code listed won't work on NT4. But looking at this
particular code, it doesn't appear to use anything Windows 2000+
specific. But it's worth watching out for if NT4 support is wanted.

Is there a decision on which platforms should be supported, other than
it's NT4+?

//Magnus