Re: [dpdk-dev] [RFC v2] bus/auxiliary: introduce auxiliary bus

["Xueming(Steven) Li" <xuemingl@nvidia.com>]

> -----Original Message-----
> From: Wang, Haiyue <haiyue.wang@intel.com>
> Sent: Tuesday, June 8, 2021 4:41 PM
> To: NBU-Contact-Thomas Monjalon <thomas@monjalon.net>; Xueming(Steven) Li <xuemingl@nvidia.com>
> Cc: dev@dpdk.org; Parav Pandit <parav@nvidia.com>; Ray Kinsella <mdr@ashroe.eu>; andrew.rybchenko@oktetlabs.ru
> Subject: RE: [dpdk-dev] [RFC v2] bus/auxiliary: introduce auxiliary bus
> 
> Hi Andrew,
> 
> > -----Original Message-----
> > From: Thomas Monjalon <thomas@monjalon.net>
> > Sent: Tuesday, June 8, 2021 15:53
> > To: Xueming Li <xuemingl@nvidia.com>
> > Cc: dev@dpdk.org; Parav Pandit <parav@nvidia.com>; Ray Kinsella
> > <mdr@ashroe.eu>; Wang, Haiyue <haiyue.wang@intel.com>;
> > andrew.rybchenko@oktetlabs.ru
> > Subject: Re: [dpdk-dev] [RFC v2] bus/auxiliary: introduce auxiliary
> > bus
> >
> 
> 
> > > +
> > > +/**
> > > + * Match function for the driver to decide if device can be handled.
> > > + *
> > > + * @param name
> > > + *   Pointer to the auxiliary device name.
> > > + * @return
> > > + *   Whether the driver can handle the auxiliary device.
> > > + */
> > > +typedef bool(*rte_auxiliary_match_t) (const char *name);
> >
> > I disagree with the earlier comment asking for typedef pointer (based
> > on one of my patches).
> > Actually Andrew's suggestion makes sense:
> > 	http://mails.dpdk.org/archives/dev/2021-June/210665.html
> >
> 
> I didn't get the error, but the same warning, I missed something ? ;-)
> 
> 1. w/o pointer
> 
> #include <stdio.h>
> 
> typedef int (gpu_malloc_t)(void *dev, size_t size, void **ptr);
> 
> static gpu_malloc_t *mlx5_gpu_malloc_fn;
> 
> static int
> mlx5_gpu_malloc(size_t size, void **ptr) {
> 	return 0;
> }
> 
> int main()
> {
> 	mlx5_gpu_malloc_fn = mlx5_gpu_malloc;
> 
> 	mlx5_gpu_malloc_fn(NULL, 0, NULL);
> 
> 	return 0;
> }
> 
> 
> gcc -Wall fun.c
> fun.c: In function 'main':
> fun.c:15:21: warning: assignment to 'int (*)(void *, size_t,  void **)' {aka 'int (*)(void *, long unsigned int,  void **)'} from incompatible
> pointer type 'int (*)(size_t,  void **)' {aka 'int (*)(long unsigned int,  void **)'} [-Wincompatible-pointer-types]
>    15 |  mlx5_gpu_malloc_fn = mlx5_gpu_malloc;
>       |                     ^
> 
> 
> 
> 2. w pointer
> 
> #include <stdio.h>
> 
> typedef int (*gpu_malloc_t)(void *dev, size_t size, void **ptr);
> 
> static gpu_malloc_t mlx5_gpu_malloc_fn;
> 
> static int
> mlx5_gpu_malloc(size_t size, void **ptr) {
> 	return 0;
> }
> 
> int main()
> {
> 	mlx5_gpu_malloc_fn = mlx5_gpu_malloc;
> 
> 	mlx5_gpu_malloc_fn(NULL, 0, NULL);
> 
> 	return 0;
> }
> 
> gcc -Wall fun.c
> fun.c: In function 'main':
> fun.c:15:21: warning: assignment to 'gpu_malloc_t' {aka 'int (*)(void *, long unsigned int,  void **)'} from incompatible pointer type
> 'int (*)(size_t,  void **)' {aka 'int (*)(long unsigned int,  void **)'} [-Wincompatible-pointer-types]
>    15 |  mlx5_gpu_malloc_fn = mlx5_gpu_malloc;
>       |                     ^
> 
> >

I don't mind as both suggestion works.
From usage perspective, it's straightforward for anybody to see an pointer field and know "that's a pointer to a something".