[PATCH] lib: make sg_pool explicitly non-modular

[PATCH] lib: make sg_pool explicitly non-modular

[Paul Gortmaker]

The recently added Kconfig controlling compilation of this code is:

lib/Kconfig:config SG_POOL
lib/Kconfig:    def_bool n

...meaning that it currently is not being built as a module by anyone.

Lets remove the modular code that is essentially orphaned, so that
when reading the driver there is no doubt it is builtin-only.

Since module_init translates to device_initcall in the non-modular
case, the init ordering remains unchanged with this commit.  However
one might want to consider moving it to subsys_initcall if it is to
be ready ahead of SCSI drivers wanting this and using device_initcall.

Cc: Christoph Hellwig <hch@lst.de>
Cc: Ming Lin <ming.l@ssi.samsung.com>
Cc: Sagi Grimberg <sagi@grimberg.me>
Cc: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
---
 lib/sg_pool.c | 17 ++---------------
 1 file changed, 2 insertions(+), 15 deletions(-)

diff --git a/lib/sg_pool.c b/lib/sg_pool.c
index 6dd30615a201..e2cf548b9610 100644
--- a/lib/sg_pool.c
+++ b/lib/sg_pool.c
@@ -1,4 +1,4 @@
-#include <linux/module.h>
+#include <linux/init.h>
 #include <linux/scatterlist.h>
 #include <linux/mempool.h>
 #include <linux/slab.h>
@@ -156,17 +156,4 @@ cleanup_sdb:
 
 	return -ENOMEM;
 }
-
-static __exit void sg_pool_exit(void)
-{
-	int i;
-
-	for (i = 0; i < SG_MEMPOOL_NR; i++) {
-		struct sg_pool *sgp = sg_pools + i;
-		mempool_destroy(sgp->pool);
-		kmem_cache_destroy(sgp->slab);
-	}
-}
-
-module_init(sg_pool_init);
-module_exit(sg_pool_exit);
+device_initcall(sg_pool_init);
-- 
2.8.0

Re: [PATCH] lib: make sg_pool explicitly non-modular

[Ming Lin]

On Wed, 2016-04-20 at 15:13 -0400, Paul Gortmaker wrote:
> The recently added Kconfig controlling compilation of this code is:
> 
> lib/Kconfig:config SG_POOL
> lib/Kconfig:    def_bool n
> 
> ...meaning that it currently is not being built as a module by anyone.
> 
> Lets remove the modular code that is essentially orphaned, so that
> when reading the driver there is no doubt it is builtin-only.
> 
> Since module_init translates to device_initcall in the non-modular
> case, the init ordering remains unchanged with this commit.  However
> one might want to consider moving it to subsys_initcall if it is to
> be ready ahead of SCSI drivers wanting this and using device_initcall.
> 
> Cc: Christoph Hellwig <hch@lst.de>
> Cc: Ming Lin <ming.l@ssi.samsung.com>
> Cc: Sagi Grimberg <sagi@grimberg.me>
> Cc: Martin K. Petersen <martin.petersen@oracle.com>
> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
> ---
>  lib/sg_pool.c | 17 ++---------------
>  1 file changed, 2 insertions(+), 15 deletions(-)
> 
> diff --git a/lib/sg_pool.c b/lib/sg_pool.c
> index 6dd30615a201..e2cf548b9610 100644
> --- a/lib/sg_pool.c
> +++ b/lib/sg_pool.c
> @@ -1,4 +1,4 @@
> -#include <linux/module.h>
> +#include <linux/init.h>
>  #include <linux/scatterlist.h>
>  #include <linux/mempool.h>
>  #include <linux/slab.h>
> @@ -156,17 +156,4 @@ cleanup_sdb:
>  
>  	return -ENOMEM;
>  }
> -
> -static __exit void sg_pool_exit(void)
> -{
> -	int i;
> -
> -	for (i = 0; i < SG_MEMPOOL_NR; i++) {
> -		struct sg_pool *sgp = sg_pools + i;
> -		mempool_destroy(sgp->pool);
> -		kmem_cache_destroy(sgp->slab);
> -	}
> -}
> -
> -module_init(sg_pool_init);
> -module_exit(sg_pool_exit);
> +device_initcall(sg_pool_init);

For SCSI it's OK because always CONFIG_SCSI=y

But we may have a kernel .config with !CONFIG_SCSI and other
non-block-device driver may use this sg_pool.

So the .config will have CONFIG_SG_POOL=m

Re: [PATCH] lib: make sg_pool explicitly non-modular

[Paul Gortmaker]

[Re: [PATCH] lib: make sg_pool explicitly non-modular] On 20/04/2016 (Wed 13:28) Ming Lin wrote:

> On Wed, 2016-04-20 at 15:13 -0400, Paul Gortmaker wrote:
> > The recently added Kconfig controlling compilation of this code is:
> > 
> > lib/Kconfig:config SG_POOL
> > lib/Kconfig:    def_bool n
> > 
> > ...meaning that it currently is not being built as a module by anyone.
> > 
> > Lets remove the modular code that is essentially orphaned, so that
> > when reading the driver there is no doubt it is builtin-only.
> > 
> > Since module_init translates to device_initcall in the non-modular
> > case, the init ordering remains unchanged with this commit.  However
> > one might want to consider moving it to subsys_initcall if it is to
> > be ready ahead of SCSI drivers wanting this and using device_initcall.
> > 
> > Cc: Christoph Hellwig <hch@lst.de>
> > Cc: Ming Lin <ming.l@ssi.samsung.com>
> > Cc: Sagi Grimberg <sagi@grimberg.me>
> > Cc: Martin K. Petersen <martin.petersen@oracle.com>
> > Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
> > ---
> >  lib/sg_pool.c | 17 ++---------------
> >  1 file changed, 2 insertions(+), 15 deletions(-)
> > 
> > diff --git a/lib/sg_pool.c b/lib/sg_pool.c
> > index 6dd30615a201..e2cf548b9610 100644
> > --- a/lib/sg_pool.c
> > +++ b/lib/sg_pool.c
> > @@ -1,4 +1,4 @@
> > -#include <linux/module.h>
> > +#include <linux/init.h>
> >  #include <linux/scatterlist.h>
> >  #include <linux/mempool.h>
> >  #include <linux/slab.h>
> > @@ -156,17 +156,4 @@ cleanup_sdb:
> >  
> >  	return -ENOMEM;
> >  }
> > -
> > -static __exit void sg_pool_exit(void)
> > -{
> > -	int i;
> > -
> > -	for (i = 0; i < SG_MEMPOOL_NR; i++) {
> > -		struct sg_pool *sgp = sg_pools + i;
> > -		mempool_destroy(sgp->pool);
> > -		kmem_cache_destroy(sgp->slab);
> > -	}
> > -}
> > -
> > -module_init(sg_pool_init);
> > -module_exit(sg_pool_exit);
> > +device_initcall(sg_pool_init);
> 
> For SCSI it's OK because always CONFIG_SCSI=y
> 
> But we may have a kernel .config with !CONFIG_SCSI and other
> non-block-device driver may use this sg_pool.
> 
> So the .config will have CONFIG_SG_POOL=m

That is impossible currently, since as per above, the variable is bool
and not tristate.  Did you mean to make it tristate?

Paul.
--

> 
> 
> 
> 

Re: [PATCH] lib: make sg_pool explicitly non-modular

[Ming Lin]

On Wed, 2016-04-20 at 23:02 -0400, Paul Gortmaker wrote:
> 
> > 
> > So the .config will have CONFIG_SG_POOL=m
> 
> That is impossible currently, since as per above, the variable is
> bool
> and not tristate.  Did you mean to make it tristate?

I didn't notice that.

Yes, it should be tristate.