u-boot/arch/arm/cpu/armv8/cache.S
Pierre-Clément Tosi 37479e65a3 armv8/cache.S: Triple with single instruction
Replace the current 2-instruction 2-step tripling code by a
corresponding single instruction leveraging ARMv8-A's "flexible second
operand as a register with optional shift". This has the added benefit
(albeit arguably negligible) of reducing the final code size.

Fix the comment as the tripled cache level is placed in x12, not x0.

Signed-off-by: Pierre-Clément Tosi <ptosi@google.com>
2021-09-23 08:55:06 -04:00

259 lines
5.7 KiB
ArmAsm

/* SPDX-License-Identifier: GPL-2.0+ */
/*
* (C) Copyright 2013
* David Feng <fenghua@phytium.com.cn>
*
* This file is based on sample code from ARMv8 ARM.
*/
#include <asm-offsets.h>
#include <config.h>
#include <asm/macro.h>
#include <asm/system.h>
#include <linux/linkage.h>
/*
* void __asm_dcache_level(level)
*
* flush or invalidate one level cache.
*
* x0: cache level
* x1: 0 clean & invalidate, 1 invalidate only
* x2~x9: clobbered
*/
.pushsection .text.__asm_dcache_level, "ax"
ENTRY(__asm_dcache_level)
lsl x12, x0, #1
msr csselr_el1, x12 /* select cache level */
isb /* sync change of cssidr_el1 */
mrs x6, ccsidr_el1 /* read the new cssidr_el1 */
ubfx x2, x6, #0, #3 /* x2 <- log2(cache line size)-4 */
ubfx x3, x6, #3, #10 /* x3 <- number of cache ways - 1 */
ubfx x4, x6, #13, #15 /* x4 <- number of cache sets - 1 */
add x2, x2, #4 /* x2 <- log2(cache line size) */
clz w5, w3 /* bit position of #ways */
/* x12 <- cache level << 1 */
/* x2 <- line length offset */
/* x3 <- number of cache ways - 1 */
/* x4 <- number of cache sets - 1 */
/* x5 <- bit position of #ways */
loop_set:
mov x6, x3 /* x6 <- working copy of #ways */
loop_way:
lsl x7, x6, x5
orr x9, x12, x7 /* map way and level to cisw value */
lsl x7, x4, x2
orr x9, x9, x7 /* map set number to cisw value */
tbz w1, #0, 1f
dc isw, x9
b 2f
1: dc cisw, x9 /* clean & invalidate by set/way */
2: subs x6, x6, #1 /* decrement the way */
b.ge loop_way
subs x4, x4, #1 /* decrement the set */
b.ge loop_set
ret
ENDPROC(__asm_dcache_level)
.popsection
/*
* void __asm_flush_dcache_all(int invalidate_only)
*
* x0: 0 clean & invalidate, 1 invalidate only
*
* flush or invalidate all data cache by SET/WAY.
*/
.pushsection .text.__asm_dcache_all, "ax"
ENTRY(__asm_dcache_all)
mov x1, x0
dsb sy
mrs x10, clidr_el1 /* read clidr_el1 */
ubfx x11, x10, #24, #3 /* x11 <- loc */
cbz x11, finished /* if loc is 0, exit */
mov x15, lr
mov x0, #0 /* start flush at cache level 0 */
/* x0 <- cache level */
/* x10 <- clidr_el1 */
/* x11 <- loc */
/* x15 <- return address */
loop_level:
add x12, x0, x0, lsl #1 /* x12 <- tripled cache level */
lsr x12, x10, x12
and x12, x12, #7 /* x12 <- cache type */
cmp x12, #2
b.lt skip /* skip if no cache or icache */
bl __asm_dcache_level /* x1 = 0 flush, 1 invalidate */
skip:
add x0, x0, #1 /* increment cache level */
cmp x11, x0
b.gt loop_level
mov x0, #0
msr csselr_el1, x0 /* restore csselr_el1 */
dsb sy
isb
mov lr, x15
finished:
ret
ENDPROC(__asm_dcache_all)
.popsection
.pushsection .text.__asm_flush_dcache_all, "ax"
ENTRY(__asm_flush_dcache_all)
mov x0, #0
b __asm_dcache_all
ENDPROC(__asm_flush_dcache_all)
.popsection
.pushsection .text.__asm_invalidate_dcache_all, "ax"
ENTRY(__asm_invalidate_dcache_all)
mov x0, #0x1
b __asm_dcache_all
ENDPROC(__asm_invalidate_dcache_all)
.popsection
/*
* void __asm_flush_dcache_range(start, end)
*
* clean & invalidate data cache in the range
*
* x0: start address
* x1: end address
*/
.pushsection .text.__asm_flush_dcache_range, "ax"
ENTRY(__asm_flush_dcache_range)
mrs x3, ctr_el0
ubfx x3, x3, #16, #4
mov x2, #4
lsl x2, x2, x3 /* cache line size */
/* x2 <- minimal cache line size in cache system */
sub x3, x2, #1
bic x0, x0, x3
1: dc civac, x0 /* clean & invalidate data or unified cache */
add x0, x0, x2
cmp x0, x1
b.lo 1b
dsb sy
ret
ENDPROC(__asm_flush_dcache_range)
.popsection
/*
* void __asm_invalidate_dcache_range(start, end)
*
* invalidate data cache in the range
*
* x0: start address
* x1: end address
*/
.pushsection .text.__asm_invalidate_dcache_range, "ax"
ENTRY(__asm_invalidate_dcache_range)
mrs x3, ctr_el0
ubfx x3, x3, #16, #4
mov x2, #4
lsl x2, x2, x3 /* cache line size */
/* x2 <- minimal cache line size in cache system */
sub x3, x2, #1
bic x0, x0, x3
1: dc ivac, x0 /* invalidate data or unified cache */
add x0, x0, x2
cmp x0, x1
b.lo 1b
dsb sy
ret
ENDPROC(__asm_invalidate_dcache_range)
.popsection
/*
* void __asm_invalidate_icache_all(void)
*
* invalidate all tlb entries.
*/
.pushsection .text.__asm_invalidate_icache_all, "ax"
ENTRY(__asm_invalidate_icache_all)
ic ialluis
isb sy
ret
ENDPROC(__asm_invalidate_icache_all)
.popsection
.pushsection .text.__asm_invalidate_l3_dcache, "ax"
WEAK(__asm_invalidate_l3_dcache)
mov x0, #0 /* return status as success */
ret
ENDPROC(__asm_invalidate_l3_dcache)
.popsection
.pushsection .text.__asm_flush_l3_dcache, "ax"
WEAK(__asm_flush_l3_dcache)
mov x0, #0 /* return status as success */
ret
ENDPROC(__asm_flush_l3_dcache)
.popsection
.pushsection .text.__asm_invalidate_l3_icache, "ax"
WEAK(__asm_invalidate_l3_icache)
mov x0, #0 /* return status as success */
ret
ENDPROC(__asm_invalidate_l3_icache)
.popsection
/*
* void __asm_switch_ttbr(ulong new_ttbr)
*
* Safely switches to a new page table.
*/
.pushsection .text.__asm_switch_ttbr, "ax"
ENTRY(__asm_switch_ttbr)
/* x2 = SCTLR (alive throghout the function) */
switch_el x4, 3f, 2f, 1f
3: mrs x2, sctlr_el3
b 0f
2: mrs x2, sctlr_el2
b 0f
1: mrs x2, sctlr_el1
0:
/* Unset CR_M | CR_C | CR_I from SCTLR to disable all caches */
movn x1, #(CR_M | CR_C | CR_I)
and x1, x2, x1
switch_el x4, 3f, 2f, 1f
3: msr sctlr_el3, x1
b 0f
2: msr sctlr_el2, x1
b 0f
1: msr sctlr_el1, x1
0: isb
/* This call only clobbers x30 (lr) and x9 (unused) */
mov x3, x30
bl __asm_invalidate_tlb_all
/* From here on we're running safely with caches disabled */
/* Set TTBR to our first argument */
switch_el x4, 3f, 2f, 1f
3: msr ttbr0_el3, x0
b 0f
2: msr ttbr0_el2, x0
b 0f
1: msr ttbr0_el1, x0
0: isb
/* Restore original SCTLR and thus enable caches again */
switch_el x4, 3f, 2f, 1f
3: msr sctlr_el3, x2
b 0f
2: msr sctlr_el2, x2
b 0f
1: msr sctlr_el1, x2
0: isb
ret x3
ENDPROC(__asm_switch_ttbr)
.popsection