@@ -31,12 +31,12 @@
*/
typedef unsigned long pte_t;
typedef unsigned long pmd_t;
-typedef unsigned long pgd_t[2];
+typedef struct { unsigned long pgd[2]; } pgd_t;
typedef unsigned long pgprot_t;
#define pte_val(x) (x)
#define pmd_val(x) (x)
-#define pgd_val(x) ((x)[0])
+#define pgd_val(x) ((x).pgd[0])
#define pgprot_val(x) (x)
#define __pte(x) (x)
@@ -24,6 +24,9 @@
typedef u32 pteval_t;
typedef u32 pmdval_t;
+typedef struct { pmdval_t pgd[2]; } pgd_t;
+#define pgd_val(x) ((x).pgd[0])
+
#undef STRICT_MM_TYPECHECKS
#ifdef STRICT_MM_TYPECHECKS
@@ -32,12 +35,10 @@ typedef u32 pmdval_t;
*/
typedef struct { pteval_t pte; } pte_t;
typedef struct { pmdval_t pmd; } pmd_t;
-typedef struct { pmdval_t pgd[2]; } pgd_t;
typedef struct { pteval_t pgprot; } pgprot_t;
#define pte_val(x) ((x).pte)
#define pmd_val(x) ((x).pmd)
-#define pgd_val(x) ((x).pgd[0])
#define pgprot_val(x) ((x).pgprot)
#define __pte(x) ((pte_t) { (x) } )
@@ -50,12 +51,10 @@ typedef struct { pteval_t pgprot; } pgprot_t;
*/
typedef pteval_t pte_t;
typedef pmdval_t pmd_t;
-typedef pmdval_t pgd_t[2];
typedef pteval_t pgprot_t;
#define pte_val(x) (x)
#define pmd_val(x) (x)
-#define pgd_val(x) ((x)[0])
#define pgprot_val(x) (x)
#define __pte(x) (x)
I ran into build errors on ARM after Willy's newly added generic TRANSPARENT_HUGEPAGE_PUD support. We don't support this feature on ARM at all, but the patch causes a build error anyway: In file included from ../kernel/memremap.c:17:0: ../include/linux/pfn_t.h:108:7: error: 'pud_mkdevmap' declared as function returning an array pud_t pud_mkdevmap(pud_t pud); We don't use a PUD on ARM, so pud_t is defined as pmd_t, which in turn is defined as typedef unsigned long pgd_t[2]; on NOMMU and on 2-level MMU configurations. There is an (unused) other definition using a struct around the array, which happens to work fine here. There is a comment in the file about the fact the other version is "easier on the compiler", and I've traced that version back to linux-2.1.80 when ARM support was first merged back in 1998. It's probably a safe assumption that this is no longer necessary: The same logic existed in asm-i386 at the time but was removed a year later in 2.3.23pre3. The STRICT_MM_TYPECHECKS logic also ended up getting copied into other architectures, which use them in various ways: * both defined, but using strict: arch/alpha/include/asm/page.h:#define STRICT_MM_TYPECHECKS arch/sparc/include/asm/page_64.h:#define STRICT_MM_TYPECHECKS arch/ia64/include/asm/page.h:# define STRICT_MM_TYPECHECKS arch/parisc/include/asm/page.h:#define STRICT_MM_TYPECHECKS * both defined, but using non-strict: arch/arc/include/asm/page.h:#undef STRICT_MM_TYPECHECKS arch/arm/include/asm/pgtable-2level-types.h:#undef STRICT_MM_TYPECHECKS arch/arm/include/asm/pgtable-3level-types.h:#undef STRICT_MM_TYPECHECKS arch/arm64/include/asm/pgtable-types.h:#undef STRICT_MM_TYPECHECKS arch/sparc/include/asm/page_32.h:/* #define STRICT_MM_TYPECHECKS */ arch/unicore32/include/asm/page.h:#undef STRICT_MM_TYPECHECKS * Kconfig option: arch/powerpc/Kconfig.debug:config STRICT_MM_TYPECHECKS default n Everthing else uses only the strict definitions. For the moment, this only does the minimal change to fix things on ARM in linux-next. At least in case of sparc32, passing structs to functions is traditionally inefficient and the non-strict version should be kept, whether we change any of the other architectures depends on further testing. Signed-off-by: Arnd Bergmann <arnd@arndb.de> Fixes: a27da20ed50e ("mm: add support for PUD-sized transparent hugepages") Link: http://www.spinics.net/lists/linux-mm/msg101404.html --- arch/arm/include/asm/page-nommu.h | 4 ++-- arch/arm/include/asm/pgtable-2level-types.h | 7 +++---- 2 files changed, 5 insertions(+), 6 deletions(-) -- 2.7.0