kernel_optimize_test/arch/arm/mm/proc-macros.S

/*
 * We need constants.h for:
 *  VMA_VM_MM
 *  VMA_VM_FLAGS
 *  VM_EXEC
 */
#include <asm/asm-offsets.h>
#include <asm/thread_info.h>

/*
 * vma_vm_mm - get mm pointer from vma pointer (vma->vm_mm)
 */
	.macro	vma_vm_mm, rd, rn
	ldr	\rd, [\rn, #VMA_VM_MM]
	.endm

/*
 * vma_vm_flags - get vma->vm_flags
 */
	.macro	vma_vm_flags, rd, rn
	ldr	\rd, [\rn, #VMA_VM_FLAGS]
	.endm

	.macro	tsk_mm, rd, rn
	ldr	\rd, [\rn, #TI_TASK]
	ldr	\rd, [\rd, #TSK_ACTIVE_MM]
	.endm

/*
 * act_mm - get current->active_mm
 */
	.macro	act_mm, rd
	bic	\rd, sp, #8128
	bic	\rd, \rd, #63
	ldr	\rd, [\rd, #TI_TASK]
	ldr	\rd, [\rd, #TSK_ACTIVE_MM]
	.endm

/*
 * mmid - get context id from mm pointer (mm->context.id)
 */
	.macro	mmid, rd, rn
	ldr	\rd, [\rn, #MM_CONTEXT_ID]
	.endm

/*
 * mask_asid - mask the ASID from the context ID
 */
	.macro	asid, rd, rn
	and	\rd, \rn, #255
	.endm

	.macro	crval, clear, mmuset, ucset
#ifdef CONFIG_MMU
	.word	\clear
	.word	\mmuset
#else
	.word	\clear
	.word	\ucset
#endif
	.endm

/*
 * cache_line_size - get the cache line size from the CSIDR register
 * (available on ARMv7+). It assumes that the CSSR register was configured
 * to access the L1 data cache CSIDR.
 */
	.macro	dcache_line_size, reg, tmp
	mrc	p15, 1, \tmp, c0, c0, 0		@ read CSIDR
	and	\tmp, \tmp, #7			@ cache line size encoding
	mov	\reg, #16			@ size offset
	mov	\reg, \reg, lsl \tmp		@ actual cache line size
	.endm


/*
 * Sanity check the PTE configuration for the code below - which makes
 * certain assumptions about how these bits are layed out.
 */
#if L_PTE_SHARED != PTE_EXT_SHARED
#error PTE shared bit mismatch
#endif
#if L_PTE_BUFFERABLE != PTE_BUFFERABLE
#error PTE bufferable bit mismatch
#endif
#if L_PTE_CACHEABLE != PTE_CACHEABLE
#error PTE cacheable bit mismatch
#endif
#if (L_PTE_EXEC+L_PTE_USER+L_PTE_WRITE+L_PTE_DIRTY+L_PTE_YOUNG+\
     L_PTE_FILE+L_PTE_PRESENT) > L_PTE_SHARED
#error Invalid Linux PTE bit settings
#endif

/*
 * The ARMv6 and ARMv7 set_pte_ext translation function.
 *
 * Permission translation:
 *  YUWD  APX AP1 AP0	SVC	User
 *  0xxx   0   0   0	no acc	no acc
 *  100x   1   0   1	r/o	no acc
 *  10x0   1   0   1	r/o	no acc
 *  1011   0   0   1	r/w	no acc
 *  110x   0   1   0	r/w	r/o
 *  11x0   0   1   0	r/w	r/o
 *  1111   0   1   1	r/w	r/w
 */
	.macro	armv6_mt_table pfx
\pfx\()_mt_table:
	.long	0x00						@ L_PTE_MT_UNCACHED
	.long	PTE_EXT_TEX(1)					@ L_PTE_MT_BUFFERABLE
	.long	PTE_CACHEABLE					@ L_PTE_MT_WRITETHROUGH
	.long	PTE_CACHEABLE | PTE_BUFFERABLE			@ L_PTE_MT_WRITEBACK
	.long	PTE_BUFFERABLE					@ L_PTE_MT_DEV_SHARED
	.long	0x00						@ unused
	.long	0x00						@ L_PTE_MT_MINICACHE (not present)
	.long	PTE_EXT_TEX(1) | PTE_CACHEABLE | PTE_BUFFERABLE	@ L_PTE_MT_WRITEALLOC
	.long	0x00						@ unused
	.long	PTE_EXT_TEX(1)					@ L_PTE_MT_DEV_WC
	.long	0x00						@ unused
	.long	PTE_CACHEABLE | PTE_BUFFERABLE			@ L_PTE_MT_DEV_CACHED
	.long	PTE_EXT_TEX(2)					@ L_PTE_MT_DEV_NONSHARED
	.long	0x00						@ L_PTE_MT_DEV_IXP2000
	.long	0x00						@ unused
	.long	0x00						@ unused
	.endm

	.macro	armv6_set_pte_ext pfx
	str	r1, [r0], #-2048		@ linux version

	bic	r3, r1, #0x000003fc
	bic	r3, r3, #PTE_TYPE_MASK
	orr	r3, r3, r2
	orr	r3, r3, #PTE_EXT_AP0 | 2

	adr	ip, \pfx\()_mt_table
	and	r2, r1, #L_PTE_MT_MASK
	ldr	r2, [ip, r2]

	tst	r1, #L_PTE_WRITE
	tstne	r1, #L_PTE_DIRTY
	orreq	r3, r3, #PTE_EXT_APX

	tst	r1, #L_PTE_USER
	orrne	r3, r3, #PTE_EXT_AP1
	tstne	r3, #PTE_EXT_APX
	bicne	r3, r3, #PTE_EXT_APX | PTE_EXT_AP0

	tst	r1, #L_PTE_EXEC
	orreq	r3, r3, #PTE_EXT_XN

	orr	r3, r3, r2

	tst	r1, #L_PTE_YOUNG
	tstne	r1, #L_PTE_PRESENT
	moveq	r3, #0

	str	r3, [r0]
	mcr	p15, 0, r0, c7, c10, 1		@ flush_pte
	.endm


/*
 * The ARMv3, ARMv4 and ARMv5 set_pte_ext translation function,
 * covering most CPUs except Xscale and Xscale 3.
 *
 * Permission translation:
 *  YUWD   AP	SVC	User
 *  0xxx  0x00	no acc	no acc
 *  100x  0x00	r/o	no acc
 *  10x0  0x00	r/o	no acc
 *  1011  0x55	r/w	no acc
 *  110x  0xaa	r/w	r/o
 *  11x0  0xaa	r/w	r/o
 *  1111  0xff	r/w	r/w
 */
	.macro	armv3_set_pte_ext wc_disable=1
	str	r1, [r0], #-2048		@ linux version

	eor	r3, r1, #L_PTE_PRESENT | L_PTE_YOUNG | L_PTE_WRITE | L_PTE_DIRTY

	bic	r2, r1, #PTE_SMALL_AP_MASK	@ keep C, B bits
	bic	r2, r2, #PTE_TYPE_MASK
	orr	r2, r2, #PTE_TYPE_SMALL

	tst	r3, #L_PTE_USER			@ user?
	orrne	r2, r2, #PTE_SMALL_AP_URO_SRW

	tst	r3, #L_PTE_WRITE | L_PTE_DIRTY	@ write and dirty?
	orreq	r2, r2, #PTE_SMALL_AP_UNO_SRW

	tst	r3, #L_PTE_PRESENT | L_PTE_YOUNG	@ present and young?
	movne	r2, #0

	.if	\wc_disable
#ifdef CONFIG_CPU_DCACHE_WRITETHROUGH
	tst	r2, #PTE_CACHEABLE
	bicne	r2, r2, #PTE_BUFFERABLE
#endif
	.endif
	str	r2, [r0]			@ hardware version
	.endm


/*
 * Xscale set_pte_ext translation, split into two halves to cope
 * with work-arounds.  r3 must be preserved by code between these
 * two macros.
 *
 * Permission translation:
 *  YUWD  AP	SVC	User
 *  0xxx  00	no acc	no acc
 *  100x  00	r/o	no acc
 *  10x0  00	r/o	no acc
 *  1011  01	r/w	no acc
 *  110x  10	r/w	r/o
 *  11x0  10	r/w	r/o
 *  1111  11	r/w	r/w
 */
	.macro	xscale_set_pte_ext_prologue
	str	r1, [r0], #-2048		@ linux version

	eor	r3, r1, #L_PTE_PRESENT | L_PTE_YOUNG | L_PTE_WRITE | L_PTE_DIRTY

	bic	r2, r1, #PTE_SMALL_AP_MASK	@ keep C, B bits
	orr	r2, r2, #PTE_TYPE_EXT		@ extended page

	tst	r3, #L_PTE_USER			@ user?
	orrne	r2, r2, #PTE_EXT_AP_URO_SRW	@ yes -> user r/o, system r/w

	tst	r3, #L_PTE_WRITE | L_PTE_DIRTY	@ write and dirty?
	orreq	r2, r2, #PTE_EXT_AP_UNO_SRW	@ yes -> user n/a, system r/w
						@ combined with user -> user r/w
	.endm

	.macro	xscale_set_pte_ext_epilogue
	tst	r3, #L_PTE_PRESENT | L_PTE_YOUNG	@ present and young?
	movne	r2, #0				@ no -> fault

	str	r2, [r0]			@ hardware version
	mov	ip, #0
	mcr	p15, 0, r0, c7, c10, 1		@ clean L1 D line
	mcr	p15, 0, ip, c7, c10, 4		@ data write barrier
	.endm