ACPICA: Utilities: Add power of two rounding support
authorLv Zheng <lv.zheng@intel.com>
Wed, 28 Dec 2016 07:28:29 +0000 (15:28 +0800)
committerRafael J. Wysocki <rafael.j.wysocki@intel.com>
Mon, 2 Jan 2017 22:18:38 +0000 (23:18 +0100)
ACPICA commit cbb0294649cbd7e8bd6107e4329461a6a7a0d967

This patch adds power of two rounding support up to 32 bits.

The result of the shift operations rearching to the boundary of the cpu
word is unpredicatable, so 64-bit roundings are not supported in order to
make sure no rounded shift-overs.

This support may not be performance friendly, so the APIs might be
overridden by the hosts implementations with ACPI_USE_NATIVE_BIT_FINDER
defined.

Link: https://github.com/acpica/acpica/commit/cbb02946
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
drivers/acpi/acpica/acmacros.h

index a3b95431b7c5cb09e3a135d3e072f7cc4685bdc8..6537cbd17bd86504639aeedec0dbf0120794e238 100644 (file)
 
 /* Generic (power-of-two) rounding */
 
+#ifndef ACPI_USE_NATIVE_BIT_FINDER
+
+#define __ACPI_FIND_LAST_BIT_2(a, r)        ((((u8)  (a)) & 0x02) ? (r)+1 : (r))
+#define __ACPI_FIND_LAST_BIT_4(a, r)        ((((u8)  (a)) & 0x0C) ? \
+                                                                                        __ACPI_FIND_LAST_BIT_2  ((a)>>2,  (r)+2) : \
+                                                                                        __ACPI_FIND_LAST_BIT_2  ((a), (r)))
+#define __ACPI_FIND_LAST_BIT_8(a, r)        ((((u8)  (a)) & 0xF0) ? \
+                                                                                        __ACPI_FIND_LAST_BIT_4  ((a)>>4,  (r)+4) : \
+                                                                                        __ACPI_FIND_LAST_BIT_4  ((a), (r)))
+#define __ACPI_FIND_LAST_BIT_16(a, r)       ((((u16) (a)) & 0xFF00) ? \
+                                                                                        __ACPI_FIND_LAST_BIT_8  ((a)>>8,  (r)+8) : \
+                                                                                        __ACPI_FIND_LAST_BIT_8  ((a), (r)))
+#define __ACPI_FIND_LAST_BIT_32(a, r)       ((((u32) (a)) & 0xFFFF0000) ? \
+                                                                                        __ACPI_FIND_LAST_BIT_16 ((a)>>16, (r)+16) : \
+                                                                                        __ACPI_FIND_LAST_BIT_16 ((a), (r)))
+#define __ACPI_FIND_LAST_BIT_64(a, r)       ((((u64) (a)) & 0xFFFFFFFF00000000) ? \
+                                                                                        __ACPI_FIND_LAST_BIT_32 ((a)>>32, (r)+32) : \
+                                                                                        __ACPI_FIND_LAST_BIT_32 ((a), (r)))
+
+#define ACPI_FIND_LAST_BIT_8(a)             ((a) ? __ACPI_FIND_LAST_BIT_8 (a, 1) : 0)
+#define ACPI_FIND_LAST_BIT_16(a)            ((a) ? __ACPI_FIND_LAST_BIT_16 (a, 1) : 0)
+#define ACPI_FIND_LAST_BIT_32(a)            ((a) ? __ACPI_FIND_LAST_BIT_32 (a, 1) : 0)
+#define ACPI_FIND_LAST_BIT_64(a)            ((a) ? __ACPI_FIND_LAST_BIT_64 (a, 1) : 0)
+
+#define __ACPI_FIND_FIRST_BIT_2(a, r)       ((((u8) (a)) & 0x01) ? (r) : (r)+1)
+#define __ACPI_FIND_FIRST_BIT_4(a, r)       ((((u8) (a)) & 0x03) ? \
+                                                                                        __ACPI_FIND_FIRST_BIT_2  ((a), (r)) : \
+                                                                                        __ACPI_FIND_FIRST_BIT_2  ((a)>>2, (r)+2))
+#define __ACPI_FIND_FIRST_BIT_8(a, r)       ((((u8) (a)) & 0x0F) ? \
+                                                                                        __ACPI_FIND_FIRST_BIT_4  ((a), (r)) : \
+                                                                                        __ACPI_FIND_FIRST_BIT_4  ((a)>>4, (r)+4))
+#define __ACPI_FIND_FIRST_BIT_16(a, r)      ((((u16) (a)) & 0x00FF) ? \
+                                                                                        __ACPI_FIND_FIRST_BIT_8  ((a), (r)) : \
+                                                                                        __ACPI_FIND_FIRST_BIT_8  ((a)>>8, (r)+8))
+#define __ACPI_FIND_FIRST_BIT_32(a, r)      ((((u32) (a)) & 0x0000FFFF) ? \
+                                                                                        __ACPI_FIND_FIRST_BIT_16 ((a), (r)) : \
+                                                                                        __ACPI_FIND_FIRST_BIT_16 ((a)>>16, (r)+16))
+#define __ACPI_FIND_FIRST_BIT_64(a, r)      ((((u64) (a)) & 0x00000000FFFFFFFF) ? \
+                                                                                        __ACPI_FIND_FIRST_BIT_32 ((a), (r)) : \
+                                                                                        __ACPI_FIND_FIRST_BIT_32 ((a)>>32, (r)+32))
+
+#define ACPI_FIND_FIRST_BIT_8(a)            ((a) ? __ACPI_FIND_FIRST_BIT_8 (a, 1) : 0)
+#define ACPI_FIND_FIRST_BIT_16(a)           ((a) ? __ACPI_FIND_FIRST_BIT_16 (a, 1) : 0)
+#define ACPI_FIND_FIRST_BIT_32(a)           ((a) ? __ACPI_FIND_FIRST_BIT_32 (a, 1) : 0)
+#define ACPI_FIND_FIRST_BIT_64(a)           ((a) ? __ACPI_FIND_FIRST_BIT_64 (a, 1) : 0)
+
+#endif                         /* ACPI_USE_NATIVE_BIT_FINDER */
+
+#define ACPI_ROUND_UP_POWER_OF_TWO_8(a)     ((u8) \
+                                                                                       (((u16) 1) <<  ACPI_FIND_LAST_BIT_8  ((a)  - 1)))
+#define ACPI_ROUND_DOWN_POWER_OF_TWO_8(a)   ((u8) \
+                                                                                       (((u16) 1) << (ACPI_FIND_LAST_BIT_8  ((a)) - 1)))
+#define ACPI_ROUND_UP_POWER_OF_TWO_16(a)    ((u16) \
+                                                                                       (((u32) 1) <<  ACPI_FIND_LAST_BIT_16 ((a)  - 1)))
+#define ACPI_ROUND_DOWN_POWER_OF_TWO_16(a)  ((u16) \
+                                                                                       (((u32) 1) << (ACPI_FIND_LAST_BIT_16 ((a)) - 1)))
+#define ACPI_ROUND_UP_POWER_OF_TWO_32(a)    ((u32) \
+                                                                                       (((u64) 1) <<  ACPI_FIND_LAST_BIT_32 ((a)  - 1)))
+#define ACPI_ROUND_DOWN_POWER_OF_TWO_32(a)  ((u32) \
+                                                                                       (((u64) 1) << (ACPI_FIND_LAST_BIT_32 ((a)) - 1)))
 #define ACPI_IS_ALIGNED(a, s)               (((a) & ((s) - 1)) == 0)
 #define ACPI_IS_POWER_OF_TWO(a)             ACPI_IS_ALIGNED(a, a)