2015-02-04 11:09:05

by Nicholas Mc Guire

[permalink] [raw]
Subject: [PATCH] staging: rtl8188eu: core: switch with redundant cases

A few redundant switch cases as well as a redundant if/else
within one of the cases was consolidated to a single call.
The cases are intentionally retained for documentation purposes.

Signed-off-by: Nicholas Mc Guire <[email protected]>
---
case WIFI_REASSOCREQ,WIFI_PROBEREQ,WIFI_BEACON,WIFI_ACTION all
have the same effect - notably the also for WIFI_PROBEREQ where
the if/else is executing the same function.

These redundant cases could all be dropped and consolidated into
the default but probably it is better for documentation/readability
to leave them in the switch/case explicitly.

Patch was only compile-tested for x86_64_defconfig + CONFIG_STAGING=y
CONFIG_R8188EU=m

Patch is against 3.19.0-rc7 (localversion-next is -next-20150204)

drivers/staging/rtl8188eu/core/rtw_mlme_ext.c | 9 ---------
1 file changed, 9 deletions(-)

diff --git a/drivers/staging/rtl8188eu/core/rtw_mlme_ext.c b/drivers/staging/rtl8188eu/core/rtw_mlme_ext.c
index 28918201..cd12dd7 100644
--- a/drivers/staging/rtl8188eu/core/rtw_mlme_ext.c
+++ b/drivers/staging/rtl8188eu/core/rtw_mlme_ext.c
@@ -484,17 +484,8 @@ void mgt_dispatcher(struct adapter *padapter, struct recv_frame *precv_frame)
/* fall through */
case WIFI_ASSOCREQ:
case WIFI_REASSOCREQ:
- _mgt_dispatcher(padapter, ptable, precv_frame);
- break;
case WIFI_PROBEREQ:
- if (check_fwstate(pmlmepriv, WIFI_AP_STATE))
- _mgt_dispatcher(padapter, ptable, precv_frame);
- else
- _mgt_dispatcher(padapter, ptable, precv_frame);
- break;
case WIFI_BEACON:
- _mgt_dispatcher(padapter, ptable, precv_frame);
- break;
case WIFI_ACTION:
_mgt_dispatcher(padapter, ptable, precv_frame);
break;
--
1.7.10.4


2015-02-04 11:42:47

by Dan Carpenter

[permalink] [raw]
Subject: Re: [PATCH] staging: rtl8188eu: core: switch with redundant cases

On Wed, Feb 04, 2015 at 06:04:38AM -0500, Nicholas Mc Guire wrote:
> A few redundant switch cases as well as a redundant if/else
> within one of the cases was consolidated to a single call.
> The cases are intentionally retained for documentation purposes.

This statement is not clear. It obviously causes a bug if you just
start deleting case statements.

>
> Signed-off-by: Nicholas Mc Guire <[email protected]>
> ---
> case WIFI_REASSOCREQ,WIFI_PROBEREQ,WIFI_BEACON,WIFI_ACTION all
> have the same effect - notably the also for WIFI_PROBEREQ where
> the if/else is executing the same function.
>
> These redundant cases could all be dropped and consolidated into
> the default but probably it is better for documentation/readability
> to leave them in the switch/case explicitly.

Oh. This explains what you meant. Stop putting this information below
the cut off, it's annoying.

regards,
dan carpenter

2015-02-04 11:43:57

by Dan Carpenter

[permalink] [raw]
Subject: Re: [PATCH] staging: rtl8188eu: core: switch with redundant cases

Btw, what tool are you using to find these?

regards,
dan carpenter

2015-02-04 12:08:08

by Rasmus Villemoes

[permalink] [raw]
Subject: Re: [PATCH] staging: rtl8188eu: core: switch with redundant cases

On Wed, Feb 04 2015, Nicholas Mc Guire <[email protected]> wrote:

> A few redundant switch cases as well as a redundant if/else
> within one of the cases was consolidated to a single call.
> The cases are intentionally retained for documentation purposes.
>
[...]
> diff --git a/drivers/staging/rtl8188eu/core/rtw_mlme_ext.c b/drivers/staging/rtl8188eu/core/rtw_mlme_ext.c
> index 28918201..cd12dd7 100644
> --- a/drivers/staging/rtl8188eu/core/rtw_mlme_ext.c
> +++ b/drivers/staging/rtl8188eu/core/rtw_mlme_ext.c
> @@ -484,17 +484,8 @@ void mgt_dispatcher(struct adapter *padapter, struct recv_frame *precv_frame)
> /* fall through */
> case WIFI_ASSOCREQ:
> case WIFI_REASSOCREQ:
> - _mgt_dispatcher(padapter, ptable, precv_frame);
> - break;
> case WIFI_PROBEREQ:
> - if (check_fwstate(pmlmepriv, WIFI_AP_STATE))
> - _mgt_dispatcher(padapter, ptable, precv_frame);
> - else
> - _mgt_dispatcher(padapter, ptable, precv_frame);

It is highly unlikely that a function called check_fwstate has side
effects, but it might be nice checking that and making a note in the
commit log.

Rasmus

2015-02-04 13:16:37

by Nicholas Mc Guire

[permalink] [raw]
Subject: Re: [PATCH] staging: rtl8188eu: core: switch with redundant cases

On Wed, 04 Feb 2015, Dan Carpenter wrote:

> Btw, what tool are you using to find these?
>
working on a set of coccinell scripts - they are not yet
really clean - but this one is simply:


virtual context
virtual patch
virtual org
virtual report

@assign@
position p;
statement S1;
@@

<+...
* if@p (...) S1 else S1
...+>

@script:python@
p << assign.p;
@@

print "%s:%s WARNING: condition with no effect" % (p[0].file,p[0].line)



I guess this is not wildly impressive - but it seems to be effective
in finding a lot of broken code. The rate of false postivs has been
supprisingly low (very few case of intentional if==else like in
./fs/kernfs/file.c:661 - which is nicely documented)

thx!
hofrat