Dan McDonald wrote:
> Hello.
>
> I *know* that one is suppoed to have one open PKCS#11 session per slot and
> per thread if you have a multithreaded PKCS#11 app. What's less clear to me
> is if:
>
>
Its been a long time since I worked on the *implementation* of PKCS#11,
but...
> - C_Initialize() is done per-thread or prior to thread creation.
> (I'd suspect once prior to threads, but I'm not sure, hence I'm
> asking.)
>
I think once per app. I suspect doing it more than once might be
incredibly destructive -- I believe part of the intention here is that
C_Initialize can be used to initialize global locks, and such.
> - What about C_GetSlotList()? My inituition says this is safe either
> way, but there are no clear docs.
>
I think its safe either way.
And for the record, I don't recall any rule about one session per thread
in the PKCS#11 specs. I *do* think that it would be incorrect for an MT
app to issue multiple simultaneous operations against a single session,
however. That said, I also have a hard time imagining a situation where
you'd use multiple threads with just a single session.
> And if someone says, "Go read PKCS#11 v2.20", I will shoot you. :)
>
I won't say it, you already have. :-)
-- Garrett
> Seriously, I was hoping to avoid reading that tome, but I will begrudgingly.
>
> Thanks,
> Dan
> _______________________________________________
> crypto-discuss mailing list
> crypto-discuss at opensolaris.org
> http://mail.opensolaris.org/mailman/listinfo/crypto-discuss
>
