Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1350695 > unrolled thread

[PATCH] cxgbit: fix dma_addr_t printk format

Started byArnd Bergmann <arnd@arndb.de>
First post2016-03-05 01:10 +0100
Last post2016-03-07 15:20 +0100
Articles 5 — 3 participants

Back to article view | Back to linux.kernel


Contents

  [PATCH] cxgbit: fix dma_addr_t printk format Arnd Bergmann <arnd@arndb.de> - 2016-03-05 01:10 +0100
    Re: [PATCH] cxgbit: fix dma_addr_t printk format Joe Perches <joe@perches.com> - 2016-03-05 01:30 +0100
      Re: [PATCH] cxgbit: fix dma_addr_t printk format Arnd Bergmann <arnd@arndb.de> - 2016-03-05 01:40 +0100
        Re: [PATCH] cxgbit: fix dma_addr_t printk format Joe Perches <joe@perches.com> - 2016-03-05 03:30 +0100
    Re: [PATCH] cxgbit: fix dma_addr_t printk format Varun Prakash <varun@chelsio.com> - 2016-03-07 15:20 +0100

#1350695 — [PATCH] cxgbit: fix dma_addr_t printk format

FromArnd Bergmann <arnd@arndb.de>
Date2016-03-05 01:10 +0100
Subject[PATCH] cxgbit: fix dma_addr_t printk format
Message-ID<r97Xc-6lT-7@gated-at.bofh.it>
The newly added driver prints a dma_addr_t using the %llx format string,
but that is wrong on most 32-bit architectures:

drivers/target/iscsi/cxgbit/cxgbit_ddp.c: In function 'cxgbit_dump_sgl':
drivers/target/iscsi/cxgbit/cxgbit_ddp.c:180:10: error: format '%llx' expects argument of type 'long long unsigned int', but argument 8 has type 'dma_addr_t {aka unsigned int}' [-Werror=format=]
   pr_info("\t%d/%u, 0x%p: len %u, off %u, pg 0x%p, dma 0x%llx, %u\n",

Unfortunately, we can't use the %pad format string here because
we are not printing an lvalue, so we have to add a cast to u64, which
matches the format string on all architectures.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Fixes: c49aa56e556d ("cxgbit: add cxgbit_ddp.c")
---
 drivers/target/iscsi/cxgbit/cxgbit_ddp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/target/iscsi/cxgbit/cxgbit_ddp.c b/drivers/target/iscsi/cxgbit/cxgbit_ddp.c
index 07e2bc86d0df..d667bc88e21d 100644
--- a/drivers/target/iscsi/cxgbit/cxgbit_ddp.c
+++ b/drivers/target/iscsi/cxgbit/cxgbit_ddp.c
@@ -179,7 +179,7 @@ cxgbit_dump_sgl(const char *cap, struct scatterlist *sgl, int nents)
 	for_each_sg(sgl, sg, nents, i)
 		pr_info("\t%d/%u, 0x%p: len %u, off %u, pg 0x%p, dma 0x%llx, %u\n",
 			i, nents, sg, sg->length, sg->offset, sg_page(sg),
-			sg_dma_address(sg), sg_dma_len(sg));
+			(u64)sg_dma_address(sg), sg_dma_len(sg));
 }
 
 static int cxgbit_ddp_sgl_check(struct scatterlist *sgl, int nents)
-- 
2.7.0

[toc] | [next] | [standalone]


#1350702

FromJoe Perches <joe@perches.com>
Date2016-03-05 01:30 +0100
Message-ID<r98gy-6uA-11@gated-at.bofh.it>
In reply to#1350695
On Sat, 2016-03-05 at 01:04 +0100, Arnd Bergmann wrote:
> The newly added driver prints a dma_addr_t using the %llx format string,
> but that is wrong on most 32-bit architectures:
> 
> drivers/target/iscsi/cxgbit/cxgbit_ddp.c: In function 'cxgbit_dump_sgl':
> drivers/target/iscsi/cxgbit/cxgbit_ddp.c:180:10: error: format '%llx' expects argument of type 'long long unsigned int', but argument 8 has type 'dma_addr_t {aka unsigned int}' [-Werror=format=]
>    pr_info("\t%d/%u, 0x%p: len %u, off %u, pg 0x%p, dma 0x%llx, %u\n",
> 
> Unfortunately, we can't use the %pad format string here because
> we are not printing an lvalue, so we have to add a cast to u64, which
> matches the format string on all architectures.
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> Fixes: c49aa56e556d ("cxgbit: add cxgbit_ddp.c")
> ---
>  drivers/target/iscsi/cxgbit/cxgbit_ddp.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/target/iscsi/cxgbit/cxgbit_ddp.c b/drivers/target/iscsi/cxgbit/cxgbit_ddp.c
> index 07e2bc86d0df..d667bc88e21d 100644
> --- a/drivers/target/iscsi/cxgbit/cxgbit_ddp.c
> +++ b/drivers/target/iscsi/cxgbit/cxgbit_ddp.c
> @@ -179,7 +179,7 @@ cxgbit_dump_sgl(const char *cap, struct scatterlist *sgl, int nents)
>  	for_each_sg(sgl, sg, nents, i)
>  		pr_info("\t%d/%u, 0x%p: len %u, off %u, pg 0x%p, dma 0x%llx, %u\n",
>  			i, nents, sg, sg->length, sg->offset, sg_page(sg),
> -			sg_dma_address(sg), sg_dma_len(sg));
> +			(u64)sg_dma_address(sg), sg_dma_len(sg));
>  }
>  
>  static int cxgbit_ddp_sgl_check(struct scatterlist *sgl, int nents)

You could create a temporary:

 	for_each_sg(sgl, sg, nents, i) {
		dma_addr_t addr = sg_dma_address(sg);

		pr_info("\t%d/%u, 0x%p: len %u, off %u, pg 0x%p, dma %pad, %u\n",
			i, nents, sg, sg->length, sg->offset, sg_page(sg),
			&addr, sg_dma_len(sg));
	}

> 

[toc] | [prev] | [next] | [standalone]


#1350710

FromArnd Bergmann <arnd@arndb.de>
Date2016-03-05 01:40 +0100
Message-ID<r98qd-6y8-5@gated-at.bofh.it>
In reply to#1350702
On Friday 04 March 2016 16:25:07 Joe Perches wrote:
> > 
> > diff --git a/drivers/target/iscsi/cxgbit/cxgbit_ddp.c b/drivers/target/iscsi/cxgbit/cxgbit_ddp.c
> > index 07e2bc86d0df..d667bc88e21d 100644
> > --- a/drivers/target/iscsi/cxgbit/cxgbit_ddp.c
> > +++ b/drivers/target/iscsi/cxgbit/cxgbit_ddp.c
> > @@ -179,7 +179,7 @@ cxgbit_dump_sgl(const char *cap, struct scatterlist *sgl, int nents)
> >       for_each_sg(sgl, sg, nents, i)
> >               pr_info("\t%d/%u, 0x%p: len %u, off %u, pg 0x%p, dma 0x%llx, %u\n",
> >                       i, nents, sg, sg->length, sg->offset, sg_page(sg),
> > -                     sg_dma_address(sg), sg_dma_len(sg));
> > +                     (u64)sg_dma_address(sg), sg_dma_len(sg));
> >  }
> >  
> >  static int cxgbit_ddp_sgl_check(struct scatterlist *sgl, int nents)
> 
> You could create a temporary:
> 
>         for_each_sg(sgl, sg, nents, i) {
>                 dma_addr_t addr = sg_dma_address(sg);
> 
>                 pr_info("\t%d/%u, 0x%p: len %u, off %u, pg 0x%p, dma %pad, %u\n",
>                         i, nents, sg, sg->length, sg->offset, sg_page(sg),
>                         &addr, sg_dma_len(sg));
>         }
> 

Sure, but the cast seemed nicer in this case, the result is the same.
	
	Arnd

[toc] | [prev] | [next] | [standalone]


#1350750

FromJoe Perches <joe@perches.com>
Date2016-03-05 03:30 +0100
Message-ID<r9a8G-7P5-9@gated-at.bofh.it>
In reply to#1350710
On Sat, 2016-03-05 at 01:34 +0100, Arnd Bergmann wrote:
> On Friday 04 March 2016 16:25:07 Joe Perches wrote:
> > > diff --git a/drivers/target/iscsi/cxgbit/cxgbit_ddp.c b/drivers/target/iscsi/cxgbit/cxgbit_ddp.c
[]
> > > @@ -179,7 +179,7 @@ cxgbit_dump_sgl(const char *cap, struct scatterlist *sgl, int nents)
> > >       for_each_sg(sgl, sg, nents, i)
> > >               pr_info("\t%d/%u, 0x%p: len %u, off %u, pg 0x%p, dma 0x%llx, %u\n",
> > >                       i, nents, sg, sg->length, sg->offset, sg_page(sg),
> > > -                     sg_dma_address(sg), sg_dma_len(sg));
> > > +                     (u64)sg_dma_address(sg), sg_dma_len(sg));
[]
> > You could create a temporary:


        for_each_sg(sgl, sg, nents, i) {
                dma_addr_t addr = sg_dma_address(sg);


                pr_info("\t%d/%u, 0x%p: len %u, off %u, pg 0x%p, dma %pad, %u\n",
                        i, nents, sg, sg->length, sg->offset, sg_page(sg),
                        &addr, sg_dma_len(sg));
        }

Sure, but the cast seemed nicer in this case, the result is the same.

Not quite as 0x%llx isn't always the same width and doesn't
have leading 0's like %pad

[toc] | [prev] | [next] | [standalone]


#1351621

FromVarun Prakash <varun@chelsio.com>
Date2016-03-07 15:20 +0100
Message-ID<ra4aS-3TP-17@gated-at.bofh.it>
In reply to#1350695
On Sat, Mar 05, 2016 at 01:04:41AM +0100, Arnd Bergmann wrote:
> The newly added driver prints a dma_addr_t using the %llx format string,
> but that is wrong on most 32-bit architectures:
> 
> drivers/target/iscsi/cxgbit/cxgbit_ddp.c: In function 'cxgbit_dump_sgl':
> drivers/target/iscsi/cxgbit/cxgbit_ddp.c:180:10: error: format '%llx' expects argument of type 'long long unsigned int', but argument 8 has type 'dma_addr_t {aka unsigned int}' [-Werror=format=]
>    pr_info("\t%d/%u, 0x%p: len %u, off %u, pg 0x%p, dma 0x%llx, %u\n",
> 
> Unfortunately, we can't use the %pad format string here because
> we are not printing an lvalue, so we have to add a cast to u64, which
> matches the format string on all architectures.
> 
> Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> Fixes: c49aa56e556d ("cxgbit: add cxgbit_ddp.c")
> ---
>  drivers/target/iscsi/cxgbit/cxgbit_ddp.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/target/iscsi/cxgbit/cxgbit_ddp.c b/drivers/target/iscsi/cxgbit/cxgbit_ddp.c
> index 07e2bc86d0df..d667bc88e21d 100644
> --- a/drivers/target/iscsi/cxgbit/cxgbit_ddp.c
> +++ b/drivers/target/iscsi/cxgbit/cxgbit_ddp.c
> @@ -179,7 +179,7 @@ cxgbit_dump_sgl(const char *cap, struct scatterlist *sgl, int nents)
>  	for_each_sg(sgl, sg, nents, i)
>  		pr_info("\t%d/%u, 0x%p: len %u, off %u, pg 0x%p, dma 0x%llx, %u\n",
>  			i, nents, sg, sg->length, sg->offset, sg_page(sg),
> -			sg_dma_address(sg), sg_dma_len(sg));
> +			(u64)sg_dma_address(sg), sg_dma_len(sg));
>  }
>  
>  static int cxgbit_ddp_sgl_check(struct scatterlist *sgl, int nents)

Ok, I will update the original commit in v2 series, thanks.

[toc] | [prev] | [standalone]


Back to top | Article view | linux.kernel


csiph-web