[patch] MM: Support more pagesizes for MAP_HUGETLB/SHM_HUGETLB v7fix fix fix

From: David Rientjes
Date: Mon Nov 12 2012 - 19:06:42 EST


On Fri, 9 Nov 2012, Randy Dunlap wrote:

> on x86_64:
>
> In file included from mm/mprotect.c:13:0:
> include/linux/shm.h:57:20: error: redefinition of 'do_shmat'
> include/linux/shm.h:57:20: note: previous definition of 'do_shmat' was here
> include/linux/shm.h:63:19: error: redefinition of 'is_file_shm_hugepages'
> include/linux/shm.h:63:19: note: previous definition of 'is_file_shm_hugepages' was here
> include/linux/shm.h:67:20: error: redefinition of 'exit_shm'
> include/linux/shm.h:67:20: note: previous definition of 'exit_shm' was here
>
> In file included from include/linux/hugetlb.h:16:0,
> from mm/mmap.c:23:
> include/linux/shm.h:57:20: error: redefinition of 'do_shmat'
> include/linux/shm.h:57:20: note: previous definition of 'do_shmat' was here
> include/linux/shm.h:63:19: error: redefinition of 'is_file_shm_hugepages'
> include/linux/shm.h:63:19: note: previous definition of 'is_file_shm_hugepages' was here
> include/linux/shm.h:67:20: error: redefinition of 'exit_shm'
> include/linux/shm.h:67:20: note: previous definition of 'exit_shm' was here
> make[2]: *** [mm/mprotect.o] Error 1
>

This is an elusive one, I couldn't reproduce it with master. This appears
to only happen on the akpm branch, correct?

It happens because of e37c64a9751a ("mm: support more pagesizes for
MAP_HUGETLB/SHM_HUGETLB") adds a spurious "#endif" to include/linux/shm.h
and gets reverted in master by 359078d82a20 ("Revert "mm: support more
pagesizes for MAP_HUGETLB/SHM_HUGETLB"") because of another build failure
that Stephen reports on November 8:

mm/mmap.c: In function 'SYSC_mmap_pgoff':
mm/mmap.c:1271:15: error: 'MAP_HUGE_SHIFT' undeclared (first use in this function)
mm/mmap.c:1271:15: note: each undeclared identifier is reported only once for each function it appears in
mm/mmap.c:1271:33: error: 'MAP_HUGE_MASK' undeclared (first use in this function)

I think the best bet would be to merge the following in -mm and then let
Stephen revert it in master again if his build error persists for powerpc:


mm-support-more-pagesizes-for-map_hugetlb-shm_hugetlb-v7-fix-fix-fix

Fix the build:

include/linux/shm.h:57:20: error: redefinition of 'do_shmat'
include/linux/shm.h:57:20: note: previous definition of 'do_shmat' was here
include/linux/shm.h:63:19: error: redefinition of 'is_file_shm_hugepages'
include/linux/shm.h:63:19: note: previous definition of 'is_file_shm_hugepages' was here
include/linux/shm.h:67:20: error: redefinition of 'exit_shm'
include/linux/shm.h:67:20: note: previous definition of 'exit_shm' was here

Signed-off-by: David Rientjes <rientjes@xxxxxxxxxx>
---
include/linux/shm.h | 4 ----
1 files changed, 0 insertions(+), 4 deletions(-)

diff --git a/include/linux/shm.h b/include/linux/shm.h
--- a/include/linux/shm.h
+++ b/include/linux/shm.h
@@ -23,8 +23,6 @@ struct shmid_kernel /* private to the kernel */
struct task_struct *shm_creator;
};

-#endif
-
/* shm_mode upper byte flags */
#define SHM_DEST 01000 /* segment will be destroyed on last detach */
#define SHM_LOCKED 02000 /* segment will not be swapped */
@@ -46,8 +44,6 @@ struct shmid_kernel /* private to the kernel */
#define SHM_HUGE_2MB (21 << SHM_HUGE_SHIFT)
#define SHM_HUGE_1GB (30 << SHM_HUGE_SHIFT)

-#ifdef __KERNEL__
-
#ifdef CONFIG_SYSVIPC
long do_shmat(int shmid, char __user *shmaddr, int shmflg, unsigned long *addr,
unsigned long shmlba);
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/