X-Git-Url: http://ftp.safe.ca/?a=blobdiff_plain;f=arch%2Farm%2Fmm%2Fproc-xsc3.S;h=2028f370288113507d2d5dcfa6e5599629270df5;hb=394168389c5770accf1d255fdfe45846ec121585;hp=8d32e21fe151a919630dc866205974888b624e3a;hpb=e6ed89ac9f5da16fea5111651b6de0ff0a76a5c2;p=safe%2Fjmp%2Flinux-2.6 diff --git a/arch/arm/mm/proc-xsc3.S b/arch/arm/mm/proc-xsc3.S index 8d32e21..2028f37 100644 --- a/arch/arm/mm/proc-xsc3.S +++ b/arch/arm/mm/proc-xsc3.S @@ -2,33 +2,33 @@ * linux/arch/arm/mm/proc-xsc3.S * * Original Author: Matthew Gilbert - * Current Maintainer: Deepak Saxena + * Current Maintainer: Lennert Buytenhek * * Copyright 2004 (C) Intel Corp. - * Copyright 2005 (c) MontaVista Software, Inc. + * Copyright 2005 (C) MontaVista Software, Inc. * * 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. * - * MMU functions for the Intel XScale3 Core (XSC3). The XSC3 core is an - * extension to Intel's original XScale core that adds the following + * MMU functions for the Intel XScale3 Core (XSC3). The XSC3 core is + * an extension to Intel's original XScale core that adds the following * features: * * - ARMv6 Supersections * - Low Locality Reference pages (replaces mini-cache) * - 36-bit addressing * - L2 cache - * - Cache-coherency if chipset supports it + * - Cache coherency if chipset supports it * - * Based on orignal XScale code by Nicolas Pitre + * Based on original XScale code by Nicolas Pitre. */ #include #include #include -#include -#include +#include +#include #include #include #include @@ -42,29 +42,19 @@ #define MAX_AREA_SIZE 32768 /* - * The cache line size of the I and D cache. + * The cache line size of the L1 I, L1 D and unified L2 cache. */ #define CACHELINESIZE 32 /* - * The size of the data cache. + * The size of the L1 D cache. */ #define CACHESIZE 32768 /* - * Run with L2 enabled. - */ -#define L2_CACHE_ENABLE 1 - -/* - * Enable the Branch Target Buffer (can cause crashes, see erratum #42.) - */ -#define BTB_ENABLE 0 - -/* - * This macro is used to wait for a CP15 write and is needed - * when we have to ensure that the last operation to the co-pro - * was completed before continuing with operation. + * This macro is used to wait for a CP15 write and is needed when we + * have to ensure that the last operation to the coprocessor was + * completed before continuing with operation. */ .macro cpwait_ret, lr, rd mrc p15, 0, \rd, c2, c0, 0 @ arbitrary read of cp15 @@ -73,13 +63,13 @@ .endm /* - * This macro cleans & invalidates the entire xsc3 dcache by set & way. + * This macro cleans and invalidates the entire L1 D cache. */ .macro clean_d_cache rd, rs mov \rd, #0x1f00 orr \rd, \rd, #0x00e0 -1: mcr p15, 0, \rd, c7, c14, 2 @ clean/inv set/way +1: mcr p15, 0, \rd, c7, c14, 2 @ clean/invalidate L1 D line adds \rd, \rd, #0x40000000 bcc 1b subs \rd, \rd, #0x20 @@ -124,15 +114,15 @@ ENTRY(cpu_xsc3_reset) mov r1, #PSR_F_BIT|PSR_I_BIT|SVC_MODE msr cpsr_c, r1 @ reset CPSR mrc p15, 0, r1, c1, c0, 0 @ ctrl register - bic r1, r1, #0x0086 @ ........B....CA. bic r1, r1, #0x3900 @ ..VIZ..S........ + bic r1, r1, #0x0086 @ ........B....CA. mcr p15, 0, r1, c1, c0, 0 @ ctrl register - mcr p15, 0, ip, c7, c7, 0 @ invalidate I,D caches & BTB + mcr p15, 0, ip, c7, c7, 0 @ invalidate L1 caches and BTB bic r1, r1, #0x0001 @ ...............M mcr p15, 0, r1, c1, c0, 0 @ ctrl register @ CAUTION: MMU turned off from this point. We count on the pipeline @ already containing those two last instructions to survive. - mcr p15, 0, ip, c8, c7, 0 @ invalidate I & D TLBs + mcr p15, 0, ip, c8, c7, 0 @ invalidate I and D TLBs mov pc, r0 /* @@ -144,14 +134,12 @@ ENTRY(cpu_xsc3_reset) * * XScale supports clock switching, but using idle mode support * allows external hardware to react to system state changes. - - MMG: Come back to this one. */ .align 5 ENTRY(cpu_xsc3_do_idle) mov r0, #1 - mcr p14, 0, r0, c7, c0, 0 @ Go to IDLE + mcr p14, 0, r0, c7, c0, 0 @ go to idle mov pc, lr /* ================================= CACHE ================================ */ @@ -176,9 +164,9 @@ ENTRY(xsc3_flush_kern_cache_all) __flush_whole_cache: clean_d_cache r0, r1 tst r2, #VM_EXEC - mcrne p15, 0, ip, c7, c5, 0 @ Invalidate I cache & BTB - mcrne p15, 0, ip, c7, c10, 4 @ Drain Write Buffer - mcrne p15, 0, ip, c7, c5, 4 @ Prefetch Flush + mcrne p15, 0, ip, c7, c5, 0 @ invalidate L1 I cache and BTB + mcrne p15, 0, ip, c7, c10, 4 @ data write barrier + mcrne p15, 0, ip, c7, c5, 4 @ prefetch flush mov pc, lr /* @@ -199,21 +187,21 @@ ENTRY(xsc3_flush_user_cache_range) bhs __flush_whole_cache 1: tst r2, #VM_EXEC - mcrne p15, 0, r0, c7, c5, 1 @ Invalidate I cache line - mcr p15, 0, r0, c7, c14, 1 @ Clean/invalidate D cache line + mcrne p15, 0, r0, c7, c5, 1 @ invalidate L1 I line + mcr p15, 0, r0, c7, c14, 1 @ clean/invalidate L1 D line add r0, r0, #CACHELINESIZE cmp r0, r1 blo 1b tst r2, #VM_EXEC - mcrne p15, 0, ip, c7, c5, 6 @ Invalidate BTB - mcrne p15, 0, ip, c7, c10, 4 @ Drain Write Buffer - mcrne p15, 0, ip, c7, c5, 4 @ Prefetch Flush + mcrne p15, 0, ip, c7, c5, 6 @ invalidate BTB + mcrne p15, 0, ip, c7, c10, 4 @ data write barrier + mcrne p15, 0, ip, c7, c5, 4 @ prefetch flush mov pc, lr /* * coherent_kern_range(start, end) * - * Ensure coherency between the Icache and the Dcache in the + * Ensure coherency between the I cache and the D cache in the * region described by start. If you have non-snooping * Harvard caches, you need to implement this function. * @@ -227,34 +215,34 @@ ENTRY(xsc3_coherent_kern_range) /* FALLTHROUGH */ ENTRY(xsc3_coherent_user_range) bic r0, r0, #CACHELINESIZE - 1 -1: mcr p15, 0, r0, c7, c10, 1 @ clean D entry +1: mcr p15, 0, r0, c7, c10, 1 @ clean L1 D line add r0, r0, #CACHELINESIZE cmp r0, r1 blo 1b mov r0, #0 - mcr p15, 0, r0, c7, c5, 0 @ Invalidate I cache & BTB - mcr p15, 0, r0, c7, c10, 4 @ Drain Write Buffer - mcr p15, 0, r0, c7, c5, 4 @ Prefetch Flush + mcr p15, 0, r0, c7, c5, 0 @ invalidate L1 I cache and BTB + mcr p15, 0, r0, c7, c10, 4 @ data write barrier + mcr p15, 0, r0, c7, c5, 4 @ prefetch flush mov pc, lr /* * flush_kern_dcache_page(void *page) * * Ensure no D cache aliasing occurs, either with itself or - * the I cache + * the I cache. * * - addr - page aligned address */ ENTRY(xsc3_flush_kern_dcache_page) add r1, r0, #PAGE_SZ -1: mcr p15, 0, r0, c7, c14, 1 @ Clean/Invalidate D Cache line +1: mcr p15, 0, r0, c7, c14, 1 @ clean/invalidate L1 D line add r0, r0, #CACHELINESIZE cmp r0, r1 blo 1b mov r0, #0 - mcr p15, 0, r0, c7, c5, 0 @ Invalidate I cache & BTB - mcr p15, 0, r0, c7, c10, 4 @ Drain Write Buffer - mcr p15, 0, r0, c7, c5, 4 @ Prefetch Flush + mcr p15, 0, r0, c7, c5, 0 @ invalidate L1 I cache and BTB + mcr p15, 0, r0, c7, c10, 4 @ data write barrier + mcr p15, 0, r0, c7, c5, 4 @ prefetch flush mov pc, lr /* @@ -271,17 +259,14 @@ ENTRY(xsc3_flush_kern_dcache_page) ENTRY(xsc3_dma_inv_range) tst r0, #CACHELINESIZE - 1 bic r0, r0, #CACHELINESIZE - 1 - mcrne p15, 0, r0, c7, c10, 1 @ clean L1 D entry - mcrne p15, 1, r0, c7, c11, 1 @ clean L2 D entry + mcrne p15, 0, r0, c7, c10, 1 @ clean L1 D line tst r1, #CACHELINESIZE - 1 - mcrne p15, 0, r1, c7, c10, 1 @ clean L1 D entry - mcrne p15, 1, r1, c7, c11, 1 @ clean L2 D entry -1: mcr p15, 0, r0, c7, c6, 1 @ invalidate L1 D entry - mcr p15, 1, r0, c7, c7, 1 @ Invalidate L2 D cache line + mcrne p15, 0, r1, c7, c10, 1 @ clean L1 D line +1: mcr p15, 0, r0, c7, c6, 1 @ invalidate L1 D line add r0, r0, #CACHELINESIZE cmp r0, r1 blo 1b - mcr p15, 0, r0, c7, c10, 4 @ Drain Write Buffer + mcr p15, 0, r0, c7, c10, 4 @ data write barrier mov pc, lr /* @@ -294,12 +279,11 @@ ENTRY(xsc3_dma_inv_range) */ ENTRY(xsc3_dma_clean_range) bic r0, r0, #CACHELINESIZE - 1 -1: mcr p15, 0, r0, c7, c10, 1 @ clean L1 D entry - mcr p15, 1, r0, c7, c11, 1 @ clean L2 D entry +1: mcr p15, 0, r0, c7, c10, 1 @ clean L1 D line add r0, r0, #CACHELINESIZE cmp r0, r1 blo 1b - mcr p15, 0, r0, c7, c10, 4 @ Drain Write Buffer + mcr p15, 0, r0, c7, c10, 4 @ data write barrier mov pc, lr /* @@ -312,13 +296,11 @@ ENTRY(xsc3_dma_clean_range) */ ENTRY(xsc3_dma_flush_range) bic r0, r0, #CACHELINESIZE - 1 -1: mcr p15, 0, r0, c7, c14, 1 @ Clean/invalidate L1 D cache line - mcr p15, 1, r0, c7, c11, 1 @ Clean L2 D cache line - mcr p15, 1, r0, c7, c7, 1 @ Invalidate L2 D cache line +1: mcr p15, 0, r0, c7, c14, 1 @ clean/invalidate L1 D line add r0, r0, #CACHELINESIZE cmp r0, r1 blo 1b - mcr p15, 0, r0, c7, c10, 4 @ Drain Write Buffer + mcr p15, 0, r0, c7, c10, 4 @ data write barrier mov pc, lr ENTRY(xsc3_cache_fns) @@ -333,7 +315,7 @@ ENTRY(xsc3_cache_fns) .long xsc3_dma_flush_range ENTRY(cpu_xsc3_dcache_clean_area) -1: mcr p15, 0, r0, c7, c10, 1 @ clean D entry +1: mcr p15, 0, r0, c7, c10, 1 @ clean L1 D line add r0, r0, #CACHELINESIZE subs r1, r1, #CACHELINESIZE bhi 1b @@ -351,52 +333,50 @@ ENTRY(cpu_xsc3_dcache_clean_area) .align 5 ENTRY(cpu_xsc3_switch_mm) clean_d_cache r1, r2 - mcr p15, 0, ip, c7, c5, 0 @ Invalidate I cache & BTB - mcr p15, 0, ip, c7, c10, 4 @ Drain Write Buffer - mcr p15, 0, ip, c7, c5, 4 @ Prefetch Flush -#ifdef L2_CACHE_ENABLE + mcr p15, 0, ip, c7, c5, 0 @ invalidate L1 I cache and BTB + mcr p15, 0, ip, c7, c10, 4 @ data write barrier + mcr p15, 0, ip, c7, c5, 4 @ prefetch flush orr r0, r0, #0x18 @ cache the page table in L2 -#endif mcr p15, 0, r0, c2, c0, 0 @ load page table pointer - mcr p15, 0, ip, c8, c7, 0 @ invalidate I & D TLBs + mcr p15, 0, ip, c8, c7, 0 @ invalidate I and D TLBs cpwait_ret lr, ip /* - * cpu_xsc3_set_pte(ptep, pte) + * cpu_xsc3_set_pte_ext(ptep, pte, ext) * * Set a PTE and flush it out - * */ - .align 5 -ENTRY(cpu_xsc3_set_pte) - str r1, [r0], #-2048 @ linux version - - bic r2, r1, #0xdf0 @ Keep C, B, coherency bits - orr r2, r2, #PTE_TYPE_EXT @ extended page - - eor r3, r1, #L_PTE_PRESENT | L_PTE_YOUNG | L_PTE_WRITE | L_PTE_DIRTY - - tst r3, #L_PTE_USER @ User? - orrne r2, r2, #PTE_EXT_AP_URO_SRW @ yes -> user r/o, system r/w +cpu_xsc3_mt_table: + .long 0x00 @ L_PTE_MT_UNCACHED + .long PTE_EXT_TEX(1) @ L_PTE_MT_BUFFERABLE + .long PTE_EXT_TEX(5) | PTE_CACHEABLE @ L_PTE_MT_WRITETHROUGH + .long PTE_CACHEABLE | PTE_BUFFERABLE @ L_PTE_MT_WRITEBACK + .long PTE_EXT_TEX(1) | PTE_BUFFERABLE @ L_PTE_MT_DEV_SHARED + .long 0x00 @ unused + .long 0x00 @ L_PTE_MT_MINICACHE (not present) + .long PTE_EXT_TEX(5) | PTE_CACHEABLE | PTE_BUFFERABLE @ L_PTE_MT_WRITEALLOC (not present?) + .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 @ unused + .long 0x00 @ unused + .long 0x00 @ unused - 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 - -#if L2_CACHE_ENABLE - @ If its cacheable it needs to be in L2 also. - eor ip, r1, #L_PTE_CACHEABLE - tst ip, #L_PTE_CACHEABLE - orreq r2, r2, #PTE_EXT_TEX(0x5) -#endif - - 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 D cache line mcr - mcr p15, 0, ip, c7, c10, 4 @ Drain Write Buffer + .align 5 +ENTRY(cpu_xsc3_set_pte_ext) + xscale_set_pte_ext_prologue + + tst r1, #L_PTE_SHARED @ shared? + and r1, r1, #L_PTE_MT_MASK + adr ip, cpu_xsc3_mt_table + ldr ip, [ip, r1] + orrne r2, r2, #PTE_EXT_COHERENT @ interlock: mask in coherent bit + bic r2, r2, #0x0c @ clear old C,B bits + orr r2, r2, ip + + xscale_set_pte_ext_epilogue mov pc, lr .ltorg @@ -409,40 +389,35 @@ ENTRY(cpu_xsc3_set_pte) __xsc3_setup: mov r0, #PSR_F_BIT|PSR_I_BIT|SVC_MODE msr cpsr_c, r0 - mcr p15, 0, ip, c7, c7, 0 @ invalidate I, D caches & BTB - mcr p15, 0, ip, c7, c10, 4 @ Drain Write Buffer - mcr p15, 0, ip, c7, c5, 4 @ Prefetch Flush - mcr p15, 0, ip, c8, c7, 0 @ invalidate I, D TLBs -#if L2_CACHE_ENABLE + mcr p15, 0, ip, c7, c7, 0 @ invalidate L1 caches and BTB + mcr p15, 0, ip, c7, c10, 4 @ data write barrier + mcr p15, 0, ip, c7, c5, 4 @ prefetch flush + mcr p15, 0, ip, c8, c7, 0 @ invalidate I and D TLBs orr r4, r4, #0x18 @ cache the page table in L2 -#endif mcr p15, 0, r4, c2, c0, 0 @ load page table pointer - mov r0, #1 @ Allow access to CP0 and CP13 - orr r0, r0, #1 << 13 @ Its undefined whether this - mcr p15, 0, r0, c15, c1, 0 @ affects USR or SVC modes + + mov r0, #0 @ don't allow CP access + mcr p15, 0, r0, c15, c1, 0 @ write CP access register + mrc p15, 0, r0, c1, c0, 1 @ get auxiliary control reg and r0, r0, #2 @ preserve bit P bit setting -#if L2_CACHE_ENABLE orr r0, r0, #(1 << 10) @ enable L2 for LLR cache -#endif mcr p15, 0, r0, c1, c0, 1 @ set auxiliary control reg + + adr r5, xsc3_crval + ldmia r5, {r5, r6} mrc p15, 0, r0, c1, c0, 0 @ get control register - bic r0, r0, #0x0002 @ .... .... .... ..A. - orr r0, r0, #0x0005 @ .... .... .... .C.M -#if BTB_ENABLE - bic r0, r0, #0x0200 @ .... ..R. .... .... - orr r0, r0, #0x3900 @ ..VI Z..S .... .... -#else - bic r0, r0, #0x0a00 @ .... Z.R. .... .... - orr r0, r0, #0x3100 @ ..VI ...S .... .... -#endif -#if L2_CACHE_ENABLE - orr r0, r0, #0x4000000 @ L2 enable -#endif + bic r0, r0, r5 @ ..V. ..R. .... ..A. + orr r0, r0, r6 @ ..VI Z..S .... .C.M (mmu) + @ ...I Z..S .... .... (uc) mov pc, lr .size __xsc3_setup, . - __xsc3_setup + .type xsc3_crval, #object +xsc3_crval: + crval clear=0x04002202, mmuset=0x00003905, ucset=0x00001900 + __INITDATA /* @@ -453,13 +428,14 @@ __xsc3_setup: .type xsc3_processor_functions, #object ENTRY(xsc3_processor_functions) .word v5t_early_abort + .word legacy_pabort .word cpu_xsc3_proc_init .word cpu_xsc3_proc_fin .word cpu_xsc3_reset .word cpu_xsc3_do_idle .word cpu_xsc3_dcache_clean_area .word cpu_xsc3_switch_mm - .word cpu_xsc3_set_pte + .word cpu_xsc3_set_pte_ext .size xsc3_processor_functions, . - xsc3_processor_functions .section ".rodata" @@ -476,7 +452,7 @@ cpu_elf_name: .type cpu_xsc3_name, #object cpu_xsc3_name: - .asciz "XScale-Core3" + .asciz "XScale-V3 based processor" .size cpu_xsc3_name, . - cpu_xsc3_name .align @@ -487,7 +463,14 @@ cpu_xsc3_name: __xsc3_proc_info: .long 0x69056000 .long 0xffffe000 - .long 0x00000c0e + .long PMD_TYPE_SECT | \ + PMD_SECT_BUFFERABLE | \ + PMD_SECT_CACHEABLE | \ + PMD_SECT_AP_WRITE | \ + PMD_SECT_AP_READ + .long PMD_TYPE_SECT | \ + PMD_SECT_AP_WRITE | \ + PMD_SECT_AP_READ b __xsc3_setup .long cpu_arch_name .long cpu_elf_name @@ -498,3 +481,28 @@ __xsc3_proc_info: .long xsc3_mc_user_fns .long xsc3_cache_fns .size __xsc3_proc_info, . - __xsc3_proc_info + +/* Note: PXA935 changed its implementor ID from Intel to Marvell */ + + .type __xsc3_pxa935_proc_info,#object +__xsc3_pxa935_proc_info: + .long 0x56056000 + .long 0xffffe000 + .long PMD_TYPE_SECT | \ + PMD_SECT_BUFFERABLE | \ + PMD_SECT_CACHEABLE | \ + PMD_SECT_AP_WRITE | \ + PMD_SECT_AP_READ + .long PMD_TYPE_SECT | \ + PMD_SECT_AP_WRITE | \ + PMD_SECT_AP_READ + b __xsc3_setup + .long cpu_arch_name + .long cpu_elf_name + .long HWCAP_SWP|HWCAP_HALF|HWCAP_THUMB|HWCAP_FAST_MULT|HWCAP_EDSP + .long cpu_xsc3_name + .long xsc3_processor_functions + .long v4wbi_tlb_fns + .long xsc3_mc_user_fns + .long xsc3_cache_fns + .size __xsc3_pxa935_proc_info, . - __xsc3_pxa935_proc_info