Mail Archives: cygwin-developers/1998/02/06/20:26:35
In article <199802070221 DOT SAA01950 AT rtl DOT cygnus DOT com>,
Geoffrey Noer <cygwin32-developers AT cygnus DOT com> wrote:
>Christopher Faylor wrote:
>>
>> How about exporting lock_pinfo and unlock_pinfo so ps.cc can use it?
>> Then it can lock the process table while it is scanning it and theoretically
>> avoid some occasional weird output.
>
>Hmmm... Not that we don't have other security holes, but this would
>mean that anyone could hang Cygwin32 trivially which I don't like very
>much. :-(
Wouldn't this only affect a specific user's mutex?
Nevermind. It is so trivial that if there is a hint of a security
problem, it is not worth investigating.
>I think a better solution here would be to make the process info
>available via an exported function(s?) which would take care of the
>locking internally.
>
>This may well be not worth the effort, however.
Nope. A functional interface into pinfo_list would be nice and would
remove the requirement of recompiling ps.cc when the structure changes
but... why bother?
--
http://www.bbc.com/ cgf AT bbc DOT com "Strange how unreal
VMS=>UNIX Solutions Boston Business Computing the real can be."
- Raw text -