2018-08-05 19:43:17

by Frank Werner-Krippendorf

[permalink] [raw]
Subject: [PATCH] Staging:rtl8712: Style - Removed inline block comment to fix "Statements should start on a tapstop"

Removed 5 inline comments "/*volatile*/" rtl87x_event.h, to fix
a coding style issue "Statements should start on a tabstop"
detected by checkpatch.pl script.

Signed-off-by: Frank Werner-Krippendorf <[email protected]>
---
drivers/staging/rtl8712/rtl871x_event.h | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/staging/rtl8712/rtl871x_event.h b/drivers/staging/rtl8712/rtl871x_event.h
index 5171379..2e59e8e 100644
--- a/drivers/staging/rtl8712/rtl871x_event.h
+++ b/drivers/staging/rtl8712/rtl871x_event.h
@@ -90,13 +90,13 @@ struct event_node {
unsigned char *node;
unsigned char evt_code;
unsigned short evt_sz;
- /*volatile*/ int *caller_ff_tail;
+ int *caller_ff_tail;
int caller_ff_sz;
};

struct c2hevent_queue {
- /*volatile*/ int head;
- /*volatile*/ int tail;
+ int head;
+ int tail;
struct event_node nodes[C2HEVENT_SZ];
unsigned char seq;
};
@@ -104,8 +104,8 @@ struct c2hevent_queue {
#define NETWORK_QUEUE_SZ 4

struct network_queue {
- /*volatile*/ int head;
- /*volatile*/ int tail;
+ int head;
+ int tail;
struct wlan_bssid_ex networks[NETWORK_QUEUE_SZ];
};

--
2.7.4



2018-08-08 09:24:44

by Greg Kroah-Hartman

[permalink] [raw]
Subject: Re: [PATCH] Staging:rtl8712: Style - Removed inline block comment to fix "Statements should start on a tapstop"

On Sun, Aug 05, 2018 at 09:41:06PM +0200, Frank Werner-Krippendorf wrote:
> Removed 5 inline comments "/*volatile*/" rtl87x_event.h, to fix
> a coding style issue "Statements should start on a tabstop"
> detected by checkpatch.pl script.
>
> Signed-off-by: Frank Werner-Krippendorf <[email protected]>
> ---
> drivers/staging/rtl8712/rtl871x_event.h | 10 +++++-----
> 1 file changed, 5 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/staging/rtl8712/rtl871x_event.h b/drivers/staging/rtl8712/rtl871x_event.h
> index 5171379..2e59e8e 100644
> --- a/drivers/staging/rtl8712/rtl871x_event.h
> +++ b/drivers/staging/rtl8712/rtl871x_event.h
> @@ -90,13 +90,13 @@ struct event_node {
> unsigned char *node;
> unsigned char evt_code;
> unsigned short evt_sz;
> - /*volatile*/ int *caller_ff_tail;
> + int *caller_ff_tail;
> int caller_ff_sz;
> };
>
> struct c2hevent_queue {
> - /*volatile*/ int head;
> - /*volatile*/ int tail;

Those "volatile" seem to imply that something is needed for these
variables. Please check to ensure that these are ok to remove in that
there is the proper locking for these types of things.

thanks,

greg k-h

2018-08-08 16:01:12

by Larry Finger

[permalink] [raw]
Subject: Re: [PATCH] Staging:rtl8712: Style - Removed inline block comment to fix "Statements should start on a tapstop"

On 08/08/2018 04:22 AM, Greg KH wrote:
> On Sun, Aug 05, 2018 at 09:41:06PM +0200, Frank Werner-Krippendorf wrote:
>> Removed 5 inline comments "/*volatile*/" rtl87x_event.h, to fix
>> a coding style issue "Statements should start on a tabstop"
>> detected by checkpatch.pl script.
>>
>> Signed-off-by: Frank Werner-Krippendorf <[email protected]>
>> ---
>> drivers/staging/rtl8712/rtl871x_event.h | 10 +++++-----
>> 1 file changed, 5 insertions(+), 5 deletions(-)
>>
>> diff --git a/drivers/staging/rtl8712/rtl871x_event.h b/drivers/staging/rtl8712/rtl871x_event.h
>> index 5171379..2e59e8e 100644
>> --- a/drivers/staging/rtl8712/rtl871x_event.h
>> +++ b/drivers/staging/rtl8712/rtl871x_event.h
>> @@ -90,13 +90,13 @@ struct event_node {
>> unsigned char *node;
>> unsigned char evt_code;
>> unsigned short evt_sz;
>> - /*volatile*/ int *caller_ff_tail;
>> + int *caller_ff_tail;
>> int caller_ff_sz;
>> };
>>
>> struct c2hevent_queue {
>> - /*volatile*/ int head;
>> - /*volatile*/ int tail;
>
> Those "volatile" seem to imply that something is needed for these
> variables. Please check to ensure that these are ok to remove in that
> there is the proper locking for these types of things.

Those in-line comments are there as a warning that these variables are special.
Just because a tool such as checkpatch says to remove them does not mean that
you should; however, if you want to get rid of the warnings, change the space(s)
after the "*/" into a tab. That silences the checkpatch warning without removing
the implicit warning that these variables are special.

NACK.

Keep in mind that when this driver was submitted, it ran checkpatch cleanly. The
fact that there are now warnings is due to changes in the tool. Following those
will be a never-ending task.

Larry


2018-08-08 19:24:15

by Dan Carpenter

[permalink] [raw]
Subject: Re: [PATCH] Staging:rtl8712: Style - Removed inline block comment to fix "Statements should start on a tapstop"

On Wed, Aug 08, 2018 at 10:59:26AM -0500, Larry Finger wrote:
> On 08/08/2018 04:22 AM, Greg KH wrote:
> > On Sun, Aug 05, 2018 at 09:41:06PM +0200, Frank Werner-Krippendorf wrote:
> > > Removed 5 inline comments "/*volatile*/" rtl87x_event.h, to fix
> > > a coding style issue "Statements should start on a tabstop"
> > > detected by checkpatch.pl script.
> > >
> > > Signed-off-by: Frank Werner-Krippendorf <[email protected]>
> > > ---
> > > drivers/staging/rtl8712/rtl871x_event.h | 10 +++++-----
> > > 1 file changed, 5 insertions(+), 5 deletions(-)
> > >
> > > diff --git a/drivers/staging/rtl8712/rtl871x_event.h b/drivers/staging/rtl8712/rtl871x_event.h
> > > index 5171379..2e59e8e 100644
> > > --- a/drivers/staging/rtl8712/rtl871x_event.h
> > > +++ b/drivers/staging/rtl8712/rtl871x_event.h
> > > @@ -90,13 +90,13 @@ struct event_node {
> > > unsigned char *node;
> > > unsigned char evt_code;
> > > unsigned short evt_sz;
> > > - /*volatile*/ int *caller_ff_tail;
> > > + int *caller_ff_tail;
> > > int caller_ff_sz;
> > > };
> > > struct c2hevent_queue {
> > > - /*volatile*/ int head;
> > > - /*volatile*/ int tail;
> >
> > Those "volatile" seem to imply that something is needed for these
> > variables. Please check to ensure that these are ok to remove in that
> > there is the proper locking for these types of things.
>
> Those in-line comments are there as a warning that these variables are
> special.

I assumed the comments meant that the locking was buggy...

I'm 100% serious. Normally that's what volatile means in a driver
context. Then I interpreted Greg's comments as meaning that warnings
are useful because they show us where the locking bugs are so we
shouldn't fix the warning until after we review the locking.

regards,
dan carpenter


2018-08-08 19:27:50

by Dan Carpenter

[permalink] [raw]
Subject: Re: [PATCH] Staging:rtl8712: Style - Removed inline block comment to fix "Statements should start on a tapstop"

Oh, yeah. Greg's comments were very clear.

regards,
dan carpenter


2018-08-08 19:35:53

by Dan Carpenter

[permalink] [raw]
Subject: Re: [PATCH] Staging:rtl8712: Style - Removed inline block comment to fix "Statements should start on a tapstop"

On Wed, Aug 08, 2018 at 11:22:57AM +0200, Greg KH wrote:
> On Sun, Aug 05, 2018 at 09:41:06PM +0200, Frank Werner-Krippendorf wrote:
> > Removed 5 inline comments "/*volatile*/" rtl87x_event.h, to fix
> > a coding style issue "Statements should start on a tabstop"
> > detected by checkpatch.pl script.
> >
> > Signed-off-by: Frank Werner-Krippendorf <[email protected]>
> > ---
> > drivers/staging/rtl8712/rtl871x_event.h | 10 +++++-----
> > 1 file changed, 5 insertions(+), 5 deletions(-)
> >
> > diff --git a/drivers/staging/rtl8712/rtl871x_event.h b/drivers/staging/rtl8712/rtl871x_event.h
> > index 5171379..2e59e8e 100644
> > --- a/drivers/staging/rtl8712/rtl871x_event.h
> > +++ b/drivers/staging/rtl8712/rtl871x_event.h
> > @@ -90,13 +90,13 @@ struct event_node {
> > unsigned char *node;
> > unsigned char evt_code;
> > unsigned short evt_sz;
> > - /*volatile*/ int *caller_ff_tail;
> > + int *caller_ff_tail;
> > int caller_ff_sz;
> > };
> >
> > struct c2hevent_queue {
> > - /*volatile*/ int head;
> > - /*volatile*/ int tail;
>
> Those "volatile" seem to imply that something is needed for these
> variables. Please check to ensure that these are ok to remove in that
> there is the proper locking for these types of things.
>

Good news, people! I have done the audit and there is no locking bug.
These structs can be deleted because they aren't ever used.

Frank, please can you resend? There may be more unused stuff in here
which I didn't see.

regards,
dan carpenter