Using ERR_CAST() is more reasonable and safer, When it is necessary
to convert the type of an error pointer and return it.

Signed-off-by: Yan Zhen <yanzhen@vivo.com>
---
 net/sunrpc/clnt.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index 09f29a95f2bc..8ee87311b348 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -603,7 +603,7 @@ struct rpc_clnt *rpc_create(struct rpc_create_args *args)
 
 	xprt = xprt_create_transport(&xprtargs);
 	if (IS_ERR(xprt))
-		return (struct rpc_clnt *)xprt;
+		return ERR_CAST(xprt);
 
 	/*
 	 * By default, kernel RPC client connects from a reserved port.
-- 
2.34.1

On Fri, 2024-08-30 at 09:42 +0800, Yan Zhen wrote:
> Using ERR_CAST() is more reasonable and safer, When it is necessary
> to convert the type of an error pointer and return it.

static inline void * __must_check ERR_CAST(__force const void *ptr)
{
        /* cast away the const */
        return (void *) ptr;
}

That function is literally just doing an implicit cast from whatever
pointer type it is now, to a 'const void *' and then to a 'void *',
which then gets implicitly cast to whatever type the caller is
expecting. Exactly how is that "safer" than the current explicit cast?

While it is great that ERR_CAST() exists, and I agree that it should be
preferred in newer code for the (sole (!)) reason that it documents
that we expect this to be an error, I see no reason why it is
imperative to apply that change to existing code. Particularly not as a
standalone patch.

So NACK for now.


-- 
Trond Myklebust
Linux NFS client maintainer, Hammerspace
trond.myklebust@hammerspace.com

From: Trond Myklebust <trondmy@hammerspace.com>
Date: Fri, 30 Aug 2024 02:08:09 +0000

> On Fri, 2024-08-30 at 09:42 +0800, Yan Zhen wrote:
>> Using ERR_CAST() is more reasonable and safer, When it is necessary
>> to convert the type of an error pointer and return it.
> 
> static inline void * __must_check ERR_CAST(__force const void *ptr)
> {
>         /* cast away the const */
>         return (void *) ptr;
> }
> 
> That function is literally just doing an implicit cast from whatever
> pointer type it is now, to a 'const void *' and then to a 'void *',
> which then gets implicitly cast to whatever type the caller is
> expecting. Exactly how is that "safer" than the current explicit cast?

I think we might want to reimplement ERR_CAST() using _Generic() to not
cast away const when the argument itself is const.

> 
> While it is great that ERR_CAST() exists, and I agree that it should be
> preferred in newer code for the (sole (!)) reason that it documents
> that we expect this to be an error, I see no reason why it is
> imperative to apply that change to existing code. Particularly not as a
> standalone patch.
> 
> So NACK for now.

Thanks,
Olek