Re: [PATCH 1/3] eal: add rte control thread create API

[Stephen Hemminger <stephen@networkplumber.org>]

On Tue, 6 Dec 2022 09:35:14 -0800
Tyler Retzlaff <roretzla@linux.microsoft.com> wrote:

> On Mon, Dec 05, 2022 at 01:11:16PM -0800, Stephen Hemminger wrote:
> > On Mon,  5 Dec 2022 12:24:26 -0800
> > Tyler Retzlaff <roretzla@linux.microsoft.com> wrote:
> >   
> > > Signed-off-by: Tyler Retzlaff <roretzla@linux.microsoft.com>
> > > ---
> > >  lib/eal/common/eal_common_thread.c | 93 ++++++++++++++++++++++++++++++++++----
> > >  lib/eal/include/rte_thread.h       | 29 ++++++++++++
> > >  lib/eal/version.map                |  3 ++
> > >  3 files changed, 117 insertions(+), 8 deletions(-)
> > > 
> > > diff --git a/lib/eal/common/eal_common_thread.c b/lib/eal/common/eal_common_thread.c
> > > index c5d8b43..ca85c51 100644
> > > --- a/lib/eal/common/eal_common_thread.c
> > > +++ b/lib/eal/common/eal_common_thread.c
> > > @@ -234,7 +234,10 @@ enum __rte_ctrl_thread_status {
> > >  };
> > >  
> > >  struct rte_thread_ctrl_params {
> > > -	void *(*start_routine)(void *);
> > > +	union {
> > > +		void * (*start_routine)(void *);
> > > +		rte_thread_func thread_func;
> > > +	} u;  
> > 
> > Why not just use rte_thread_func, this in internal.  
> 
> I'm not sure i completely understand your comment here. The main reason
> for using a union is to avoid dealing with casting.  Later when the
> rte_ctrl_thread_create is deprecated the union will be discarded.
> 
> No change was made in v2 but if you still think this should be addressed
> let me know.

Is it possible to do it without union or casting.
Aliasing function pointers is something that leads to long term issues.
The Linux kernel developers are actively fixing up all the casting there;
would like DPDK to adopt the same hygiene.