Re: [PATCH 1/8] verbs: Always allocate a verbs_context

[Jason Gunthorpe <jgg-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>]

On Wed, Jan 10, 2018 at 11:16:20AM +0530, Devesh Sharma wrote:

> > +/*
> > + * Allocate and initialize a context structure. This is called to create the
> > + * driver wrapper, and context_offset is the number of bytes into the wrapper
> > + * structure where the verbs_context starts.
> > + */
> > +void *_verbs_init_and_alloc_context(struct ibv_device *device, int cmd_fd,
> > +                                   size_t alloc_size,
> > +                                   struct verbs_context *context_offset)
> > +{
> > +       void *drv_context;
> > +       struct verbs_context *context;
> > +
> > +       drv_context = calloc(1, alloc_size);
> > +       if (!drv_context) {
> > +               errno = ENOMEM;
> > +               close(cmd_fd);
> > +               return NULL;
> > +       }
> > +
> > +       context = (struct verbs_context *)((uint8_t *)drv_context +
> > +                                          (uintptr_t)context_offset);
> 
> A wrapper macro would do better here?

What would we call it? It is kinda of a reverse container of

Thing is, this is the only place that does this calculation and it is
intimately tied to the definition of the
verbs_init_and_alloc_context() macro, so it really is unique and
special to this function.

To elaborate on what is happening here..

The driver calls

        cntx = verbs_init_and_alloc_context(vdev, cmd_fd, cntx, ibvctx);

Where:
        struct bnxt_re_context *cntx;

And the name 'ibvctx' is like container_of, where it refers to the
struct member inside cntx:

struct bnxt_re_context {
        struct verbs_context ibvctx;

This allows the allocation function to both return the 'struct
bnxt_re_context' and find the 'struct verbs_context' where the driver
placed it.

The alternative here is to force the driver to put the verbs_context
at the start of the struct, then eliminate the context_offset entirely
and replace it with a static assert scheme.

Do you think that is nicer?

Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html