Message ID | 20180525214548.2122779-1-arnd@arndb.de |
---|---|
State | Not Applicable |
Headers | show |
Series | procfs: use inline functions instead of macros for proc_create_single_data stub | expand |
Looks fine:
Reviewed-by: Christoph Hellwig <hch@lst.de>
On Fri, 25 May 2018 23:45:00 +0200 Arnd Bergmann <arnd@arndb.de> wrote: > The procfs interface changes caused one warning in afs for > a now unused function: > > fs/afs/proc.c:818:12: error: 'afs_proc_stats_show' defined but not used [-Werror=unused-function] > static int afs_proc_stats_show(struct seq_file *m, void *v) > > This can be avoided by using an inline function instead of a macro > to reference it, so now the compiler can silently drop the function > after seeing that there is a reference but that it is never called. > > Unfortunately, this change triggers another warning where a function > is hidden and now unexpectedly referenced: > > drivers/mtd/mtdcore.c: In function 'init_mtd': > drivers/mtd/mtdcore.c:1878:48: error: 'mtd_proc_show' undeclared (first use in this function); did you mean 'mtd_name_show'? > > It seems nicer to keep using the inline function and removing the #ifdef > here than to add an #ifdef around every single function we pass into > proc_create_single_data(), so I'm removing the #ifdef here. > > After a few hundred randconfig builds, this was the only instance > I found that caused a problem. > > Fixes: 353861cf0594 ("afs: simplify procfs code") > Fixes: 3f3942aca6da ("proc: introduce proc_create_single{,_data}") > Cc: Christoph Hellwig <hch@lst.de> > Signed-off-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Boris Brezillon <boris.brezillon@bootlin.com> > --- > drivers/mtd/mtdcore.c | 3 --- > include/linux/proc_fs.h | 10 ++++++++-- > 2 files changed, 8 insertions(+), 5 deletions(-) > > diff --git a/drivers/mtd/mtdcore.c b/drivers/mtd/mtdcore.c > index 42395df06be9..08d1e89faf9c 100644 > --- a/drivers/mtd/mtdcore.c > +++ b/drivers/mtd/mtdcore.c > @@ -1814,8 +1814,6 @@ void *mtd_kmalloc_up_to(const struct mtd_info *mtd, size_t *size) > } > EXPORT_SYMBOL_GPL(mtd_kmalloc_up_to); > > -#ifdef CONFIG_PROC_FS > - > /*====================================================================*/ > /* Support for /proc/mtd */ > > @@ -1833,7 +1831,6 @@ static int mtd_proc_show(struct seq_file *m, void *v) > mutex_unlock(&mtd_table_mutex); > return 0; > } > -#endif /* CONFIG_PROC_FS */ > > /*====================================================================*/ > /* Init code */ > diff --git a/include/linux/proc_fs.h b/include/linux/proc_fs.h > index e518352137e7..3b44c357a6e7 100644 > --- a/include/linux/proc_fs.h > +++ b/include/linux/proc_fs.h > @@ -84,8 +84,14 @@ static inline struct proc_dir_entry *proc_mkdir_mode(const char *name, > #define proc_create_seq_private(name, mode, parent, ops, size, data) ({NULL;}) > #define proc_create_seq_data(name, mode, parent, ops, data) ({NULL;}) > #define proc_create_seq(name, mode, parent, ops) ({NULL;}) > -#define proc_create_single(name, mode, parent, show) ({NULL;}) > -#define proc_create_single_data(name, mode, parent, show, data) ({NULL;}) > +static inline struct proc_dir_entry *proc_create_single_data(const char *name, > + umode_t mode, struct proc_dir_entry *parent, > + int (*show)(struct seq_file *, void *), void *data) > +{ > + return NULL; > +} > +#define proc_create_single(name, mode, parent, show) \ > + proc_create_single_data(name, mode, parent, show, NULL) > #define proc_create(name, mode, parent, proc_fops) ({NULL;}) > #define proc_create_data(name, mode, parent, proc_fops, data) ({NULL;}) >
diff --git a/drivers/mtd/mtdcore.c b/drivers/mtd/mtdcore.c index 42395df06be9..08d1e89faf9c 100644 --- a/drivers/mtd/mtdcore.c +++ b/drivers/mtd/mtdcore.c @@ -1814,8 +1814,6 @@ void *mtd_kmalloc_up_to(const struct mtd_info *mtd, size_t *size) } EXPORT_SYMBOL_GPL(mtd_kmalloc_up_to); -#ifdef CONFIG_PROC_FS - /*====================================================================*/ /* Support for /proc/mtd */ @@ -1833,7 +1831,6 @@ static int mtd_proc_show(struct seq_file *m, void *v) mutex_unlock(&mtd_table_mutex); return 0; } -#endif /* CONFIG_PROC_FS */ /*====================================================================*/ /* Init code */ diff --git a/include/linux/proc_fs.h b/include/linux/proc_fs.h index e518352137e7..3b44c357a6e7 100644 --- a/include/linux/proc_fs.h +++ b/include/linux/proc_fs.h @@ -84,8 +84,14 @@ static inline struct proc_dir_entry *proc_mkdir_mode(const char *name, #define proc_create_seq_private(name, mode, parent, ops, size, data) ({NULL;}) #define proc_create_seq_data(name, mode, parent, ops, data) ({NULL;}) #define proc_create_seq(name, mode, parent, ops) ({NULL;}) -#define proc_create_single(name, mode, parent, show) ({NULL;}) -#define proc_create_single_data(name, mode, parent, show, data) ({NULL;}) +static inline struct proc_dir_entry *proc_create_single_data(const char *name, + umode_t mode, struct proc_dir_entry *parent, + int (*show)(struct seq_file *, void *), void *data) +{ + return NULL; +} +#define proc_create_single(name, mode, parent, show) \ + proc_create_single_data(name, mode, parent, show, NULL) #define proc_create(name, mode, parent, proc_fops) ({NULL;}) #define proc_create_data(name, mode, parent, proc_fops, data) ({NULL;})
The procfs interface changes caused one warning in afs for a now unused function: fs/afs/proc.c:818:12: error: 'afs_proc_stats_show' defined but not used [-Werror=unused-function] static int afs_proc_stats_show(struct seq_file *m, void *v) This can be avoided by using an inline function instead of a macro to reference it, so now the compiler can silently drop the function after seeing that there is a reference but that it is never called. Unfortunately, this change triggers another warning where a function is hidden and now unexpectedly referenced: drivers/mtd/mtdcore.c: In function 'init_mtd': drivers/mtd/mtdcore.c:1878:48: error: 'mtd_proc_show' undeclared (first use in this function); did you mean 'mtd_name_show'? It seems nicer to keep using the inline function and removing the #ifdef here than to add an #ifdef around every single function we pass into proc_create_single_data(), so I'm removing the #ifdef here. After a few hundred randconfig builds, this was the only instance I found that caused a problem. Fixes: 353861cf0594 ("afs: simplify procfs code") Fixes: 3f3942aca6da ("proc: introduce proc_create_single{,_data}") Cc: Christoph Hellwig <hch@lst.de> Signed-off-by: Arnd Bergmann <arnd@arndb.de> --- drivers/mtd/mtdcore.c | 3 --- include/linux/proc_fs.h | 10 ++++++++-- 2 files changed, 8 insertions(+), 5 deletions(-)