RE: checkpatch.pl report about "Missing blank line after declarations" in a structure definition

[Dotan Barak <dotanb@mellanox.com>]

Hi Joe.

The patch that you mentioned solved most of the issues, thanks!

However, there is still one more warning of this type within a struct declaration.

# ./checkpatch.pl --file --no-tree ../include/linux/mlx5/driver.h

WARNING: Missing a blank line after declarations
#508: FILE: ../include/linux/mlx5/driver.h:508:
+       struct mlx5_uuar_info   uuari;
+       MLX5_DECLARE_DOORBELL_LOCK(cq_uar_lock);


<snip start>
struct mlx5_priv {
        char                    name[MLX5_MAX_NAME_LEN];
        struct mlx5_eq_table    eq_table;
        struct mlx5_uuar_info   uuari;
        MLX5_DECLARE_DOORBELL_LOCK(cq_uar_lock);

        /* pages stuff */
        struct workqueue_struct *pg_wq;
        struct rb_root          page_root;
        int                     fw_pages;
        int                     reg_pages;
        struct list_head        free_list;

<snip end>


Thanks
Dotan

> -----Original Message-----
> From: Joe Perches [mailto:joe@perches.com]
> Sent: Monday, August 04, 2014 4:36 PM
> To: Dotan Barak
> Cc: linux-kernel@vger.kernel.org
> Subject: Re: checkpatch.pl report about "Missing blank line after
> declarations" in a structure definition
> 
> On Mon, 2014-08-04 at 11:20 +0000, Dotan Barak wrote:
> > Hi Joe.
> >
> > Thanks for fixing the issue that I reported about.
> >
> > It seems that even after your fix, there is still a false warning of "Missing
> blank line after declarations".
> >
> > I executed checkpatch.pl on a header file that exists in the Linux tree:
> > # ./checkpatch.pl --file --no-tree ../include/linux/mlx5/cq.h
> >
> > And got the following warning:
> >
> > WARNING: Missing a blank line after declarations
> > #49: FILE: ../include/linux/mlx5/cq.h:49:
> > +       int                     irqn;
> > +       void (*comp)            (struct mlx5_core_cq *);
> []
> > As you can see, this is a structure definition and IMHO the warning above is
> false.
> 
> The version in Linus' tree has a defect where this test doesn't recognize
> function pointer declarations.
> 
> This is fixed in -next since June, but it takes awhile to get into Linus' tree.
> 
> https://lkml.org/lkml/2014/6/6/426
>