Re: [rfc] epoll interface change and glibc bits ...

[Jakub Jelinek <jakub@redhat.com>]

On Mon, Nov 18, 2002 at 08:15:39AM -0800, Davide Libenzi wrote:
> > > 1) epoll's event structure extension
> > >
> > > I received quite a few request to extend the event structure to have space
> > > for an opaque user data object. The eventpoll event structure will turn to
> > > be :
> > >
> > > struct epollfd {
> > > 	int fd;
> > > 	unsigned short int events, revents;
> > > 	unsigned long obj;
> >
> > Cannot this be uint64_t obj; ?
> > Have mercy with 32<->64 bit translation layers in the kernel.
> 
> Maybe this :
> 
> typedef void *epoll_obj_t;

That is as bad as unsigned long - it is different between 32-bit and 64-bit
ABIs.
Unless you handle the translation in the generic epoll code, there will
need to be wrappers around it which will kmalloc (maxelements * sizeof (struct epollfd));
do the generic epoll into that array and then translate all the structures
in that array into smaller ones and copy to userspace.

	Jakub