Undermining XGrabKeyboard

Eirik Byrkjeflot Anonsen eirik at opera.com
Wed Mar 5 00:30:58 PST 2008


pcpa at mandriva.com.br writes:

> Quoting Peter Hutterer <mailinglists at who-t.net>:
>
>> Paulo Cesar Pereira de Andrade wrote:
>>>   I wrote a patch for the Mandriva X Server that handles keyboard
>>> events using sigio, like mouse events,
>>
>> unless you wrote a lot more code than this, this will re-create the 
>> problem that X mallocs and frees memory in the signal handler.
>
>   I am not 100% sure about this, so feel free to correct me, but I believe
> libc malloc is reentrant

I expect most general mallocs are thread safe these days, but actually
making malloc signal safe is a whole other matter.  I'm not sure which
meaning of "reentrant" you meant :)

(By the way, there are more meanings to "reentrant" than those two...)

While it is certainly possible to make malloc signal safe (though I
expect that will greatly increase the likelihood of it returning
NULL), I doubt that many mallocs are designed to be signal safe.
Depending on malloc being signal safe sounds a bit optimistic to me...

eirik



More information about the xorg mailing list