kernel_optimize_test/arch/arm/mm/proc-arm940.S
Lorenzo Pieralisi 031bd879f7 ARM: mm: implement LoUIS API for cache maintenance ops
ARM v7 architecture introduced the concept of cache levels and related
control registers. New processors like A7 and A15 embed an L2 unified cache
controller that becomes part of the cache level hierarchy. Some operations in
the kernel like cpu_suspend and __cpu_disable do not require a flush of the
entire cache hierarchy to DRAM but just the cache levels belonging to the
Level of Unification Inner Shareable (LoUIS), which in most of ARM v7 systems
correspond to L1.

The current cache flushing API used in cpu_suspend and __cpu_disable,
flush_cache_all(), ends up flushing the whole cache hierarchy since for
v7 it cleans and invalidates all cache levels up to Level of Coherency
(LoC) which cripples system performance when used in hot paths like hotplug
and cpuidle.

Therefore a new kernel cache maintenance API must be added to cope with
latest ARM system requirements.

This patch adds flush_cache_louis() to the ARM kernel cache maintenance API.

This function cleans and invalidates all data cache levels up to the
Level of Unification Inner Shareable (LoUIS) and invalidates the instruction
cache for processors that support it (> v7).

This patch also creates an alias of the cache LoUIS function to flush_kern_all
for all processor versions prior to v7, so that the current cache flushing
behaviour is unchanged for those processors.

v7 cache maintenance code implements a cache LoUIS function that cleans and
invalidates the D-cache up to LoUIS and invalidates the I-cache, according
to the new API.

Reviewed-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
Reviewed-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Tested-by: Shawn Guo <shawn.guo@linaro.org>
2012-09-25 11:20:25 +01:00

377 lines
9.5 KiB
ArmAsm

/*
* linux/arch/arm/mm/arm940.S: utility functions for ARM940T
*
* Copyright (C) 2004-2006 Hyok S. Choi (hyok.choi@samsung.com)
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2 as
* published by the Free Software Foundation.
*
*/
#include <linux/linkage.h>
#include <linux/init.h>
#include <asm/assembler.h>
#include <asm/hwcap.h>
#include <asm/pgtable-hwdef.h>
#include <asm/pgtable.h>
#include <asm/ptrace.h>
#include "proc-macros.S"
/* ARM940T has a 4KB DCache comprising 256 lines of 4 words */
#define CACHE_DLINESIZE 16
#define CACHE_DSEGMENTS 4
#define CACHE_DENTRIES 64
.text
/*
* cpu_arm940_proc_init()
* cpu_arm940_switch_mm()
*
* These are not required.
*/
ENTRY(cpu_arm940_proc_init)
ENTRY(cpu_arm940_switch_mm)
mov pc, lr
/*
* cpu_arm940_proc_fin()
*/
ENTRY(cpu_arm940_proc_fin)
mrc p15, 0, r0, c1, c0, 0 @ ctrl register
bic r0, r0, #0x00001000 @ i-cache
bic r0, r0, #0x00000004 @ d-cache
mcr p15, 0, r0, c1, c0, 0 @ disable caches
mov pc, lr
/*
* cpu_arm940_reset(loc)
* Params : r0 = address to jump to
* Notes : This sets up everything for a reset
*/
.pushsection .idmap.text, "ax"
ENTRY(cpu_arm940_reset)
mov ip, #0
mcr p15, 0, ip, c7, c5, 0 @ flush I cache
mcr p15, 0, ip, c7, c6, 0 @ flush D cache
mcr p15, 0, ip, c7, c10, 4 @ drain WB
mrc p15, 0, ip, c1, c0, 0 @ ctrl register
bic ip, ip, #0x00000005 @ .............c.p
bic ip, ip, #0x00001000 @ i-cache
mcr p15, 0, ip, c1, c0, 0 @ ctrl register
mov pc, r0
ENDPROC(cpu_arm940_reset)
.popsection
/*
* cpu_arm940_do_idle()
*/
.align 5
ENTRY(cpu_arm940_do_idle)
mcr p15, 0, r0, c7, c0, 4 @ Wait for interrupt
mov pc, lr
/*
* flush_icache_all()
*
* Unconditionally clean and invalidate the entire icache.
*/
ENTRY(arm940_flush_icache_all)
mov r0, #0
mcr p15, 0, r0, c7, c5, 0 @ invalidate I cache
mov pc, lr
ENDPROC(arm940_flush_icache_all)
/*
* flush_user_cache_all()
*/
ENTRY(arm940_flush_user_cache_all)
/* FALLTHROUGH */
/*
* flush_kern_cache_all()
*
* Clean and invalidate the entire cache.
*/
ENTRY(arm940_flush_kern_cache_all)
mov r2, #VM_EXEC
/* FALLTHROUGH */
/*
* flush_user_cache_range(start, end, flags)
*
* There is no efficient way to flush a range of cache entries
* in the specified address range. Thus, flushes all.
*
* - start - start address (inclusive)
* - end - end address (exclusive)
* - flags - vm_flags describing address space
*/
ENTRY(arm940_flush_user_cache_range)
mov ip, #0
#ifdef CONFIG_CPU_DCACHE_WRITETHROUGH
mcr p15, 0, ip, c7, c6, 0 @ flush D cache
#else
mov r1, #(CACHE_DSEGMENTS - 1) << 4 @ 4 segments
1: orr r3, r1, #(CACHE_DENTRIES - 1) << 26 @ 64 entries
2: mcr p15, 0, r3, c7, c14, 2 @ clean/flush D index
subs r3, r3, #1 << 26
bcs 2b @ entries 63 to 0
subs r1, r1, #1 << 4
bcs 1b @ segments 3 to 0
#endif
tst r2, #VM_EXEC
mcrne p15, 0, ip, c7, c5, 0 @ invalidate I cache
mcrne p15, 0, ip, c7, c10, 4 @ drain WB
mov pc, lr
/*
* coherent_kern_range(start, end)
*
* Ensure coherency between the Icache and the Dcache in the
* region described by start, end. If you have non-snooping
* Harvard caches, you need to implement this function.
*
* - start - virtual start address
* - end - virtual end address
*/
ENTRY(arm940_coherent_kern_range)
/* FALLTHROUGH */
/*
* coherent_user_range(start, end)
*
* Ensure coherency between the Icache and the Dcache in the
* region described by start, end. If you have non-snooping
* Harvard caches, you need to implement this function.
*
* - start - virtual start address
* - end - virtual end address
*/
ENTRY(arm940_coherent_user_range)
/* FALLTHROUGH */
/*
* flush_kern_dcache_area(void *addr, size_t size)
*
* Ensure no D cache aliasing occurs, either with itself or
* the I cache
*
* - addr - kernel address
* - size - region size
*/
ENTRY(arm940_flush_kern_dcache_area)
mov r0, #0
mov r1, #(CACHE_DSEGMENTS - 1) << 4 @ 4 segments
1: orr r3, r1, #(CACHE_DENTRIES - 1) << 26 @ 64 entries
2: mcr p15, 0, r3, c7, c14, 2 @ clean/flush D index
subs r3, r3, #1 << 26
bcs 2b @ entries 63 to 0
subs r1, r1, #1 << 4
bcs 1b @ segments 7 to 0
mcr p15, 0, r0, c7, c5, 0 @ invalidate I cache
mcr p15, 0, r0, c7, c10, 4 @ drain WB
mov pc, lr
/*
* dma_inv_range(start, end)
*
* There is no efficient way to invalidate a specifid virtual
* address range. Thus, invalidates all.
*
* - start - virtual start address
* - end - virtual end address
*/
arm940_dma_inv_range:
mov ip, #0
mov r1, #(CACHE_DSEGMENTS - 1) << 4 @ 4 segments
1: orr r3, r1, #(CACHE_DENTRIES - 1) << 26 @ 64 entries
2: mcr p15, 0, r3, c7, c6, 2 @ flush D entry
subs r3, r3, #1 << 26
bcs 2b @ entries 63 to 0
subs r1, r1, #1 << 4
bcs 1b @ segments 7 to 0
mcr p15, 0, ip, c7, c10, 4 @ drain WB
mov pc, lr
/*
* dma_clean_range(start, end)
*
* There is no efficient way to clean a specifid virtual
* address range. Thus, cleans all.
*
* - start - virtual start address
* - end - virtual end address
*/
arm940_dma_clean_range:
ENTRY(cpu_arm940_dcache_clean_area)
mov ip, #0
#ifndef CONFIG_CPU_DCACHE_WRITETHROUGH
mov r1, #(CACHE_DSEGMENTS - 1) << 4 @ 4 segments
1: orr r3, r1, #(CACHE_DENTRIES - 1) << 26 @ 64 entries
2: mcr p15, 0, r3, c7, c10, 2 @ clean D entry
subs r3, r3, #1 << 26
bcs 2b @ entries 63 to 0
subs r1, r1, #1 << 4
bcs 1b @ segments 7 to 0
#endif
mcr p15, 0, ip, c7, c10, 4 @ drain WB
mov pc, lr
/*
* dma_flush_range(start, end)
*
* There is no efficient way to clean and invalidate a specifid
* virtual address range.
*
* - start - virtual start address
* - end - virtual end address
*/
ENTRY(arm940_dma_flush_range)
mov ip, #0
mov r1, #(CACHE_DSEGMENTS - 1) << 4 @ 4 segments
1: orr r3, r1, #(CACHE_DENTRIES - 1) << 26 @ 64 entries
2:
#ifndef CONFIG_CPU_DCACHE_WRITETHROUGH
mcr p15, 0, r3, c7, c14, 2 @ clean/flush D entry
#else
mcr p15, 0, r3, c7, c6, 2 @ invalidate D entry
#endif
subs r3, r3, #1 << 26
bcs 2b @ entries 63 to 0
subs r1, r1, #1 << 4
bcs 1b @ segments 7 to 0
mcr p15, 0, ip, c7, c10, 4 @ drain WB
mov pc, lr
/*
* dma_map_area(start, size, dir)
* - start - kernel virtual start address
* - size - size of region
* - dir - DMA direction
*/
ENTRY(arm940_dma_map_area)
add r1, r1, r0
cmp r2, #DMA_TO_DEVICE
beq arm940_dma_clean_range
bcs arm940_dma_inv_range
b arm940_dma_flush_range
ENDPROC(arm940_dma_map_area)
/*
* dma_unmap_area(start, size, dir)
* - start - kernel virtual start address
* - size - size of region
* - dir - DMA direction
*/
ENTRY(arm940_dma_unmap_area)
mov pc, lr
ENDPROC(arm940_dma_unmap_area)
.globl arm940_flush_kern_cache_louis
.equ arm940_flush_kern_cache_louis, arm940_flush_kern_cache_all
@ define struct cpu_cache_fns (see <asm/cacheflush.h> and proc-macros.S)
define_cache_functions arm940
__CPUINIT
.type __arm940_setup, #function
__arm940_setup:
mov r0, #0
mcr p15, 0, r0, c7, c5, 0 @ invalidate I cache
mcr p15, 0, r0, c7, c6, 0 @ invalidate D cache
mcr p15, 0, r0, c7, c10, 4 @ drain WB
mcr p15, 0, r0, c6, c3, 0 @ disable data area 3~7
mcr p15, 0, r0, c6, c4, 0
mcr p15, 0, r0, c6, c5, 0
mcr p15, 0, r0, c6, c6, 0
mcr p15, 0, r0, c6, c7, 0
mcr p15, 0, r0, c6, c3, 1 @ disable instruction area 3~7
mcr p15, 0, r0, c6, c4, 1
mcr p15, 0, r0, c6, c5, 1
mcr p15, 0, r0, c6, c6, 1
mcr p15, 0, r0, c6, c7, 1
mov r0, #0x0000003F @ base = 0, size = 4GB
mcr p15, 0, r0, c6, c0, 0 @ set area 0, default
mcr p15, 0, r0, c6, c0, 1
ldr r0, =(CONFIG_DRAM_BASE & 0xFFFFF000) @ base[31:12] of RAM
ldr r1, =(CONFIG_DRAM_SIZE >> 12) @ size of RAM (must be >= 4KB)
mov r2, #10 @ 11 is the minimum (4KB)
1: add r2, r2, #1 @ area size *= 2
mov r1, r1, lsr #1
bne 1b @ count not zero r-shift
orr r0, r0, r2, lsl #1 @ the area register value
orr r0, r0, #1 @ set enable bit
mcr p15, 0, r0, c6, c1, 0 @ set area 1, RAM
mcr p15, 0, r0, c6, c1, 1
ldr r0, =(CONFIG_FLASH_MEM_BASE & 0xFFFFF000) @ base[31:12] of FLASH
ldr r1, =(CONFIG_FLASH_SIZE >> 12) @ size of FLASH (must be >= 4KB)
mov r2, #10 @ 11 is the minimum (4KB)
1: add r2, r2, #1 @ area size *= 2
mov r1, r1, lsr #1
bne 1b @ count not zero r-shift
orr r0, r0, r2, lsl #1 @ the area register value
orr r0, r0, #1 @ set enable bit
mcr p15, 0, r0, c6, c2, 0 @ set area 2, ROM/FLASH
mcr p15, 0, r0, c6, c2, 1
mov r0, #0x06
mcr p15, 0, r0, c2, c0, 0 @ Region 1&2 cacheable
mcr p15, 0, r0, c2, c0, 1
#ifdef CONFIG_CPU_DCACHE_WRITETHROUGH
mov r0, #0x00 @ disable whole write buffer
#else
mov r0, #0x02 @ Region 1 write bufferred
#endif
mcr p15, 0, r0, c3, c0, 0
mov r0, #0x10000
sub r0, r0, #1 @ r0 = 0xffff
mcr p15, 0, r0, c5, c0, 0 @ all read/write access
mcr p15, 0, r0, c5, c0, 1
mrc p15, 0, r0, c1, c0 @ get control register
orr r0, r0, #0x00001000 @ I-cache
orr r0, r0, #0x00000005 @ MPU/D-cache
mov pc, lr
.size __arm940_setup, . - __arm940_setup
__INITDATA
@ define struct processor (see <asm/proc-fns.h> and proc-macros.S)
define_processor_functions arm940, dabort=nommu_early_abort, pabort=legacy_pabort, nommu=1
.section ".rodata"
string cpu_arch_name, "armv4t"
string cpu_elf_name, "v4"
string cpu_arm940_name, "ARM940T"
.align
.section ".proc.info.init", #alloc, #execinstr
.type __arm940_proc_info,#object
__arm940_proc_info:
.long 0x41009400
.long 0xff00fff0
.long 0
b __arm940_setup
.long cpu_arch_name
.long cpu_elf_name
.long HWCAP_SWP | HWCAP_HALF | HWCAP_THUMB
.long cpu_arm940_name
.long arm940_processor_functions
.long 0
.long 0
.long arm940_cache_fns
.size __arm940_proc_info, . - __arm940_proc_info