[PATCH net-next] net: socket: use BIT_MASK for MSG_*

[PATCH net-next] net: socket: use BIT_MASK for MSG_*

[menglong8.dong]

From: Menglong Dong <dong.menglong@zte.com.cn>

The bit mask for MSG_* seems a little confused here. Replace it
with BIT_MASK to make it clear to understand.

Signed-off-by: Menglong Dong <dong.menglong@zte.com.cn>
---
 include/linux/socket.h | 71 ++++++++++++++++++++++--------------------
 1 file changed, 37 insertions(+), 34 deletions(-)

diff --git a/include/linux/socket.h b/include/linux/socket.h
index 385894b4a8bb..671d31c41582 100644
--- a/include/linux/socket.h
+++ b/include/linux/socket.h
@@ -283,42 +283,45 @@ struct ucred {
    Added those for 1003.1g not all are supported yet
  */
 
-#define MSG_OOB		1
-#define MSG_PEEK	2
-#define MSG_DONTROUTE	4
-#define MSG_TRYHARD     4       /* Synonym for MSG_DONTROUTE for DECnet */
-#define MSG_CTRUNC	8
-#define MSG_PROBE	0x10	/* Do not send. Only probe path f.e. for MTU */
-#define MSG_TRUNC	0x20
-#define MSG_DONTWAIT	0x40	/* Nonblocking io		 */
-#define MSG_EOR         0x80	/* End of record */
-#define MSG_WAITALL	0x100	/* Wait for a full request */
-#define MSG_FIN         0x200
-#define MSG_SYN		0x400
-#define MSG_CONFIRM	0x800	/* Confirm path validity */
-#define MSG_RST		0x1000
-#define MSG_ERRQUEUE	0x2000	/* Fetch message from error queue */
-#define MSG_NOSIGNAL	0x4000	/* Do not generate SIGPIPE */
-#define MSG_MORE	0x8000	/* Sender will send more */
-#define MSG_WAITFORONE	0x10000	/* recvmmsg(): block until 1+ packets avail */
-#define MSG_SENDPAGE_NOPOLICY 0x10000 /* sendpage() internal : do no apply policy */
-#define MSG_SENDPAGE_NOTLAST 0x20000 /* sendpage() internal : not the last page */
-#define MSG_BATCH	0x40000 /* sendmmsg(): more messages coming */
-#define MSG_EOF         MSG_FIN
-#define MSG_NO_SHARED_FRAGS 0x80000 /* sendpage() internal : page frags are not shared */
-#define MSG_SENDPAGE_DECRYPTED	0x100000 /* sendpage() internal : page may carry
-					  * plain text and require encryption
-					  */
-
-#define MSG_ZEROCOPY	0x4000000	/* Use user data in kernel path */
-#define MSG_FASTOPEN	0x20000000	/* Send data in TCP SYN */
-#define MSG_CMSG_CLOEXEC 0x40000000	/* Set close_on_exec for file
-					   descriptor received through
-					   SCM_RIGHTS */
+#define MSG_OOB		BIT_MASK(0)
+#define MSG_PEEK	BIT_MASK(1)
+#define MSG_DONTROUTE	BIT_MASK(2)
+#define MSG_TRYHARD	BIT_MASK(2)	/* Synonym for MSG_DONTROUTE for DECnet		*/
+#define MSG_CTRUNC	BIT_MASK(3)
+#define MSG_PROBE	BIT_MASK(4)	/* Do not send. Only probe path f.e. for MTU	*/
+#define MSG_TRUNC	BIT_MASK(5)
+#define MSG_DONTWAIT	BIT_MASK(6)	/* Nonblocking io		*/
+#define MSG_EOR		BIT_MASK(7)	/* End of record		*/
+#define MSG_WAITALL	BIT_MASK(8)	/* Wait for a full request	*/
+#define MSG_FIN		BIT_MASK(9)
+#define MSG_SYN		BIT_MASK(10)
+#define MSG_CONFIRM	BIT_MASK(11)	/* Confirm path validity	*/
+#define MSG_RST		BIT_MASK(12)
+#define MSG_ERRQUEUE	BIT_MASK(13)	/* Fetch message from error queue */
+#define MSG_NOSIGNAL	BIT_MASK(14)	/* Do not generate SIGPIPE	*/
+#define MSG_MORE	BIT_MASK(15)	/* Sender will send more	*/
+#define MSG_WAITFORONE	BIT_MASK(16)	/* recvmmsg(): block until 1+ packets avail */
+#define MSG_SENDPAGE_NOPOLICY	BIT_MASK(16)	/* sendpage() internal : do no apply policy */
+#define MSG_SENDPAGE_NOTLAST	BIT_MASK(17)	/* sendpage() internal : not the last page  */
+#define MSG_BATCH	BIT_MASK(18)		/* sendmmsg(): more messages coming */
+#define MSG_EOF		MSG_FIN
+#define MSG_NO_SHARED_FRAGS	BIT_MASK(19)	/* sendpage() internal : page frags
+						 * are not shared
+						 */
+#define MSG_SENDPAGE_DECRYPTED	BIT_MASK(20)	/* sendpage() internal : page may carry
+						 * plain text and require encryption
+						 */
+
+#define MSG_ZEROCOPY	BIT_MASK(26)	/* Use user data in kernel path */
+#define MSG_FASTOPEN	BIT_MASK(29)	/* Send data in TCP SYN */
+#define MSG_CMSG_CLOEXEC	BIT_MASK(30)	/* Set close_on_exec for file
+						 * descriptor received through
+						 * SCM_RIGHTS
+						 */
 #if defined(CONFIG_COMPAT)
-#define MSG_CMSG_COMPAT	0x80000000	/* This message needs 32 bit fixups */
+#define MSG_CMSG_COMPAT	BIT_MASK(31)	/* This message needs 32 bit fixups */
 #else
-#define MSG_CMSG_COMPAT	0		/* We never have 32 bit fixups */
+#define MSG_CMSG_COMPAT	0	/* We never have 32 bit fixups */
 #endif
 
 
-- 
2.30.0

Re: [PATCH net-next] net: socket: use BIT_MASK for MSG_*

[Menglong Dong]

Hello!

On Sat, Feb 6, 2021 at 4:20 PM Andy Shevchenko
<andy.shevchenko@gmail.com> wrote:
>
>
>
> On Saturday, February 6, 2021, <menglong8.dong@gmail.com> wrote:
>>
>> From: Menglong Dong <dong.menglong@zte.com.cn>
>>
>> The bit mask for MSG_* seems a little confused here. Replace it
>> with BIT_MASK to make it clear to understand.
>
>
> It makes it more confusing if you understand the difference between BIT_MASK() and BIT(). I think you have to use the latter. And note () when referring to the function or macro.
>

I replaced BIT_MASK() with BIT() in the patch of v2, and it looks much
more tidy.
I can't figure out the difference between BIT() and BIT_MASK(), seems
the latter one more safe... isn't it?

Thanks:)
Menglong Dong

Re: [PATCH net-next] net: socket: use BIT_MASK for MSG_*

[Andy Shevchenko]

On Sun, Feb 7, 2021 at 5:29 AM Menglong Dong <menglong8.dong@gmail.com> wrote:
> On Sat, Feb 6, 2021 at 4:20 PM Andy Shevchenko
> <andy.shevchenko@gmail.com> wrote:
> > On Saturday, February 6, 2021, <menglong8.dong@gmail.com> wrote:

> > It makes it more confusing if you understand the difference between BIT_MASK() and BIT(). I think you have to use the latter. And note () when referring to the function or macro.
>
> I replaced BIT_MASK() with BIT() in the patch of v2, and it looks much
> more tidy.
> I can't figure out the difference between BIT() and BIT_MASK(), seems
> the latter one more safe... isn't it?

BIT_MASK() operates on top of (long) bitmaps when you already know the
address of the certain word you would like to change. When the
parameter is constant it will be helpful only in some (rare) cases,
where you define a (long) bit mask, which has a bit, like 65, set. For
most of the cases it's not needed. It's rare that hardware operates on
bitmaps longer than data bus width.


-- 
With Best Regards,
Andy Shevchenko