[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <YkMc9p4nd/4+2fkf@kroah.com>
Date: Tue, 29 Mar 2022 16:51:34 +0200
From: Greg KH <gregkh@...uxfoundation.org>
To: Nam Cao <namcaov@...il.com>
Cc: sudipm.mukherjee@...il.com, teddy.wang@...iconmotion.com,
linux-staging@...ts.linux.dev, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] staging: sm750fb: add necessary #include in
sm750_accel.h, sm750_cursor.h
On Fri, Mar 18, 2022 at 05:50:46PM +0100, Nam Cao wrote:
> In sm750_cursor.h, struct lynx_cursor is used without including the
> header file where this struct is defined. Similar thing is seen in
> sm750_accel.h.
>
> The module can still be compiled because there is "#include "sm750.h""
> before every "#include "sm750_accel.h"" and "#include "sm750_cursor.h"".
> However, the order of #include should not affect the ability to compile.
>
> Add "#include "sm750.h"" to sm750_accel.h and sm750_cursor.h.
>
> Signed-off-by: Nam Cao <namcaov@...il.com>
> ---
> drivers/staging/sm750fb/sm750_accel.h | 2 ++
> drivers/staging/sm750fb/sm750_cursor.h | 2 ++
> 2 files changed, 4 insertions(+)
>
> diff --git a/drivers/staging/sm750fb/sm750_accel.h b/drivers/staging/sm750fb/sm750_accel.h
> index 2c79cb730a0a..fe6ff196272c 100644
> --- a/drivers/staging/sm750fb/sm750_accel.h
> +++ b/drivers/staging/sm750fb/sm750_accel.h
> @@ -2,6 +2,8 @@
> #ifndef ACCEL_H__
> #define ACCEL_H__
>
> +#include "sm750.h"
> +
> #define HW_ROP2_COPY 0xc
> #define HW_ROP2_XOR 0x6
>
> diff --git a/drivers/staging/sm750fb/sm750_cursor.h b/drivers/staging/sm750fb/sm750_cursor.h
> index b59643dd61ed..f023b691ac0b 100644
> --- a/drivers/staging/sm750fb/sm750_cursor.h
> +++ b/drivers/staging/sm750fb/sm750_cursor.h
> @@ -2,6 +2,8 @@
> #ifndef LYNX_CURSOR_H__
> #define LYNX_CURSOR_H__
>
> +#include "sm750.h"
> +
> /* hw_cursor_xxx works for voyager,718 and 750 */
> void sm750_hw_cursor_enable(struct lynx_cursor *cursor);
> void sm750_hw_cursor_disable(struct lynx_cursor *cursor);
> --
> 2.25.1
>
>
I think the current code is fine, given that this is not actually
"fixing" anything. If you wish to unwind all of the #include mess in
this driver, that would be fine, but I doubt it really is worth it. You
really should not need more than 1 .h file for a single module, this one
is split up way too much.
thanks,
greg k-h
Powered by blists - more mailing lists