diff options
| author | Chris Metcalf <cmetcalf@tilera.com> | 2013-08-07 19:55:35 +0400 |
|---|---|---|
| committer | Chris Metcalf <cmetcalf@tilera.com> | 2013-08-14 00:26:05 +0400 |
| commit | ba02f0eb826da6dbdc5a2958ac52304ee441234f (patch) | |
| tree | 8b50a046b6dd1a4cbc1379ae43c1c673c39e8d3d /arch/tile/include/asm/unaligned.h | |
| parent | bc1a298f4e04833db4c430df59b90039f0170515 (diff) | |
| download | linux-ba02f0eb826da6dbdc5a2958ac52304ee441234f.tar.xz | |
tile: improve big-endian support
First, fix a bug in asm/unaligned.h; we need to just use the asm-generic
unaligned.h so we properly choose endian-correct flavors.
Second, keep the hv/hypervisor.h ABI fully "native" in the sense that
we don't have __BIG_ENDIAN__ ifdefs there. Instead, we use macros in
the head_NN.S assembly code to properly extract two 32-bit structure
members from a 64-bit register holding the structure.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
Diffstat (limited to 'arch/tile/include/asm/unaligned.h')
| -rw-r--r-- | arch/tile/include/asm/unaligned.h | 14 |
1 files changed, 9 insertions, 5 deletions
diff --git a/arch/tile/include/asm/unaligned.h b/arch/tile/include/asm/unaligned.h index 37dfbe598872..5a58a0d11449 100644 --- a/arch/tile/include/asm/unaligned.h +++ b/arch/tile/include/asm/unaligned.h @@ -15,11 +15,15 @@ #ifndef _ASM_TILE_UNALIGNED_H #define _ASM_TILE_UNALIGNED_H -#include <linux/unaligned/le_struct.h> -#include <linux/unaligned/be_byteshift.h> -#include <linux/unaligned/generic.h> -#define get_unaligned __get_unaligned_le -#define put_unaligned __put_unaligned_le +/* + * We could implement faster get_unaligned_[be/le]64 using the ldna + * instruction on tilegx; however, we need to either copy all of the + * other generic functions to here (which is pretty ugly) or else + * modify both the generic code and other arch code to allow arch + * specific unaligned data access functions. Given these functions + * are not often called, we'll stick with the generic version. + */ +#include <asm-generic/unaligned.h> /* * Is the kernel doing fixups of unaligned accesses? If <0, no kernel |