[PATCH libibverbs] memory: use SCNxPTR format to read uintptr_t values

[PATCH libibverbs] memory: use SCNxPTR format to read uintptr_t values

[Yann Droneaud]

SCNxPTR is the correct format to parse uintptr_t hexadecimal values,
whatever the width of uintptr_t type.

This fix a warning when building on a 32bits system.

Signed-off-by: Yann Droneaud <ydroneaud-RlY5vtjFyJ3QT0dZR+AlfA@public.gmane.org>
---
 src/memory.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/src/memory.c b/src/memory.c
index ea9e712..7d97e55 100644
--- a/src/memory.c
+++ b/src/memory.c
@@ -44,6 +44,7 @@
 #include <string.h>
 #include <dirent.h>
 #include <limits.h>
+#include <inttypes.h>
 
 #include "ibverbs.h"
 
@@ -116,7 +117,7 @@ static unsigned long get_page_size(void *base)
 		int n;
 		uintptr_t range_start, range_end;
 
-		n = sscanf(buf, "%lx-%lx", &range_start, &range_end);
+		n = sscanf(buf, "%" SCNxPTR "-%" SCNxPTR, &range_start, &range_end);
 
 		if (n < 2)
 			continue;
-- 
1.7.2.5

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH libibverbs] memory: use SCNxPTR format to read uintptr_t values

[Roland Dreier]

On Wed, Jul 11, 2012 at 8:10 AM, Yann Droneaud <ydroneaud-RlY5vtjFyJ3QT0dZR+AlfA@public.gmane.org> wrote:
> @@ -116,7 +117,7 @@ static unsigned long get_page_size(void *base)
>                 int n;
>                 uintptr_t range_start, range_end;
>
> -               n = sscanf(buf, "%lx-%lx", &range_start, &range_end);
> +               n = sscanf(buf, "%" SCNxPTR "-%" SCNxPTR, &range_start, &range_end);
>
>                 if (n < 2)
>                         continue;

Looks fine in itself, and I don't necessarily expect you to be the one
to answer, but:

 - what does /proc/<pid>/maps show when running a 32-bit process on a 64-bit
   kernel?  Should we be using uint64_t / SCNx64?  (but surely 32-bit processes
   are guaranteed to have all their mappings fit into 32 bits)

 - earlier in this function, why do we do /proc/%d/maps, getpid()?  Why doesn't
   /proc/self/maps always work?

Thanks,
  Roland
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH libibverbs] memory: use SCNxPTR format to read uintptr_t values

[Jason Gunthorpe]

On Wed, Jul 11, 2012 at 08:47:05AM -0700, Roland Dreier wrote:

> Looks fine in itself, and I don't necessarily expect you to be the one
> to answer, but:
> 
>  - what does /proc/<pid>/maps show when running a 32-bit process on a 64-bit
>    kernel?  Should we be using uint64_t / SCNx64?  (but surely 32-bit processes
>    are guaranteed to have all their mappings fit into 32 bits)

Linux 2.6.35.9 #3 Wed Jan 5 12:06:51 MST 2011 x86_64 GNU/Linux

08048000-08052000 r-xp 00000000 ca:00 32652                              /bin/cat

So PTR seems right..

Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH libibverbs] memory: use SCNxPTR format to read uintptr_t values

[Yann Droneaud]

Le mercredi 11 juillet 2012 à 08:47 -0700, Roland Dreier a écrit :
> On Wed, Jul 11, 2012 at 8:10 AM, Yann Droneaud <ydroneaud-RlY5vtjFyJ3QT0dZR+AlfA@public.gmane.org> wrote:
> > @@ -116,7 +117,7 @@ static unsigned long get_page_size(void *base)
> >                 int n;
> >                 uintptr_t range_start, range_end;
> >
> > -               n = sscanf(buf, "%lx-%lx", &range_start, &range_end);
> > +               n = sscanf(buf, "%" SCNxPTR "-%" SCNxPTR, &range_start, &range_end);
> >
> >                 if (n < 2)
> >                         continue;
> 
> Looks fine in itself, and I don't necessarily expect you to be the one
> to answer, but:
> 
>  - what does /proc/<pid>/maps show when running a 32-bit process on a 64-bit
>    kernel?  Should we be using uint64_t / SCNx64?  (but surely 32-bit processes
>    are guaranteed to have all their mappings fit into 32 bits)
> 

"maps" file is presented to 32bits process as a 32bits mapping.

>  - earlier in this function, why do we do /proc/%d/maps, getpid()?  Why doesn't
>    /proc/self/maps always work?
> 

I think /proc/self/maps would always work here.

Using explicitly the pid could be of use for the child to read its
parent mapping.

Regards

-- 
Yann Droneaud
OPTEYA


--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html