On Tue, 2005-03-15 at 08:24 -0800, Randy.Dunlap wrote:
> Evgeniy Polyakov wrote:
> > --- /tmp/empty/crypto_main.c        1970-01-01 03:00:00.000000000 +0300
> > +++ ./acrypto/crypto_main.c 2005-03-07 20:35:36.000000000 +0300
> > @@ -0,0 +1,374 @@
> > +/*
> > + *         crypto_main.c
> > + *
> > + * Copyright (c) 2004 Evgeniy Polyakov <[EMAIL PROTECTED]>
> > + * 
> > + */
> 
> > +struct crypto_session *crypto_session_alloc(struct 
> > crypto_session_initializer *ci, struct crypto_data *d)
> > +{
> > +   struct crypto_session *s;
> > +
> > +   s = crypto_session_create(ci, d);
> > +   if (!s)
> > +           return NULL;
> > +
> > +   crypto_session_add(s);
> > +
> > +   return s;
> > +}
> > +
> > +
> 
> > +EXPORT_SYMBOL(crypto_session_alloc);
> Why is this one not _GPL ??  It calls _create() and _add().

It is not allowed to control _create() and _add() methods, only call
them "atomically"
(without gap between functions where new route can be created).
So I export only that one functin as non-GPL-only for anyone
who wants to use asynchronous crypto in simple mode.
More powerfull control requires GPL.

> > +EXPORT_SYMBOL_GPL(crypto_session_create);
> > +EXPORT_SYMBOL_GPL(crypto_session_add);
> > +EXPORT_SYMBOL_GPL(crypto_session_dequeue_route);
> 
> 
-- 
        Evgeniy Polyakov

Crash is better than data corruption -- Arthur Grabowski

Attachment: signature.asc
Description: This is a digitally signed message part

Reply via email to