From 2f78eae5064728d6cd907148cfeaf8ba3e63b0ef Mon Sep 17 00:00:00 2001 From: York Sun Date: Mon, 23 Jun 2014 15:15:54 -0700 Subject: ARMv8/FSL_LSCH3: Add FSL_LSCH3 SoC Freescale LayerScape with Chassis Generation 3 is a set of SoCs with ARMv8 cores and 3rd generation of Chassis. We use different MMU setup to support memory map and cache attribute for these SoCs. MMU and cache are enabled very early to bootst performance, especially for early development on emulators. After u-boot relocates to DDR, a new MMU table with QBMan cache access is created in DDR. SMMU pagesize is set in SMMU_sACR register. Both DDR3 and DDR4 are supported. Signed-off-by: York Sun Signed-off-by: Varun Sethi Signed-off-by: Arnab Basu --- arch/arm/cpu/armv8/fsl-lsch3/cpu.c | 425 +++++++++++++++++++++++++++++++++++++ 1 file changed, 425 insertions(+) create mode 100644 arch/arm/cpu/armv8/fsl-lsch3/cpu.c (limited to 'arch/arm/cpu/armv8/fsl-lsch3/cpu.c') diff --git a/arch/arm/cpu/armv8/fsl-lsch3/cpu.c b/arch/arm/cpu/armv8/fsl-lsch3/cpu.c new file mode 100644 index 0000000000..46965f08ad --- /dev/null +++ b/arch/arm/cpu/armv8/fsl-lsch3/cpu.c @@ -0,0 +1,425 @@ +/* + * Copyright 2014 Freescale Semiconductor, Inc. + * + * SPDX-License-Identifier: GPL-2.0+ + */ + +#include +#include +#include +#include +#include +#include +#include "cpu.h" +#include "speed.h" + +DECLARE_GLOBAL_DATA_PTR; + +#ifndef CONFIG_SYS_DCACHE_OFF +/* + * To start MMU before DDR is available, we create MMU table in SRAM. + * The base address of SRAM is CONFIG_SYS_FSL_OCRAM_BASE. We use three + * levels of translation tables here to cover 40-bit address space. + * We use 4KB granule size, with 40 bits physical address, T0SZ=24 + * Level 0 IA[39], table address @0 + * Level 1 IA[31:30], table address @01000, 0x2000 + * Level 2 IA[29:21], table address @0x3000 + */ + +#define SECTION_SHIFT_L0 39UL +#define SECTION_SHIFT_L1 30UL +#define SECTION_SHIFT_L2 21UL +#define BLOCK_SIZE_L0 0x8000000000UL +#define BLOCK_SIZE_L1 (1 << SECTION_SHIFT_L1) +#define BLOCK_SIZE_L2 (1 << SECTION_SHIFT_L2) +#define CONFIG_SYS_IFC_BASE 0x30000000 +#define CONFIG_SYS_IFC_SIZE 0x10000000 +#define CONFIG_SYS_IFC_BASE2 0x500000000 +#define CONFIG_SYS_IFC_SIZE2 0x100000000 +#define TCR_EL2_PS_40BIT (2 << 16) +#define LSCH3_VA_BITS (40) +#define LSCH3_TCR (TCR_TG0_4K | \ + TCR_EL2_PS_40BIT | \ + TCR_SHARED_NON | \ + TCR_ORGN_NC | \ + TCR_IRGN_NC | \ + TCR_T0SZ(LSCH3_VA_BITS)) + +/* + * Final MMU + * Let's start from the same layout as early MMU and modify as needed. + * IFC regions will be cache-inhibit. + */ +#define FINAL_QBMAN_CACHED_MEM 0x818000000UL +#define FINAL_QBMAN_CACHED_SIZE 0x4000000 + + +static inline void early_mmu_setup(void) +{ + int el; + u64 i; + u64 section_l1t0, section_l1t1, section_l2; + u64 *level0_table = (u64 *)CONFIG_SYS_FSL_OCRAM_BASE; + u64 *level1_table_0 = (u64 *)(CONFIG_SYS_FSL_OCRAM_BASE + 0x1000); + u64 *level1_table_1 = (u64 *)(CONFIG_SYS_FSL_OCRAM_BASE + 0x2000); + u64 *level2_table = (u64 *)(CONFIG_SYS_FSL_OCRAM_BASE + 0x3000); + + + level0_table[0] = + (u64)level1_table_0 | PMD_TYPE_TABLE; + level0_table[1] = + (u64)level1_table_1 | PMD_TYPE_TABLE; + + /* + * set level 1 table 0 to cache_inhibit, covering 0 to 512GB + * set level 1 table 1 to cache enabled, covering 512GB to 1TB + * set level 2 table to cache-inhibit, covering 0 to 1GB + */ + section_l1t0 = 0; + section_l1t1 = BLOCK_SIZE_L0; + section_l2 = 0; + for (i = 0; i < 512; i++) { + set_pgtable_section(level1_table_0, i, section_l1t0, + MT_DEVICE_NGNRNE); + set_pgtable_section(level1_table_1, i, section_l1t1, + MT_NORMAL); + set_pgtable_section(level2_table, i, section_l2, + MT_DEVICE_NGNRNE); + section_l1t0 += BLOCK_SIZE_L1; + section_l1t1 += BLOCK_SIZE_L1; + section_l2 += BLOCK_SIZE_L2; + } + + level1_table_0[0] = + (u64)level2_table | PMD_TYPE_TABLE; + level1_table_0[1] = + 0x40000000 | PMD_SECT_AF | PMD_TYPE_SECT | + PMD_ATTRINDX(MT_DEVICE_NGNRNE); + level1_table_0[2] = + 0x80000000 | PMD_SECT_AF | PMD_TYPE_SECT | + PMD_ATTRINDX(MT_NORMAL); + level1_table_0[3] = + 0xc0000000 | PMD_SECT_AF | PMD_TYPE_SECT | + PMD_ATTRINDX(MT_NORMAL); + + /* Rewrite table to enable cache */ + set_pgtable_section(level2_table, + CONFIG_SYS_FSL_OCRAM_BASE >> SECTION_SHIFT_L2, + CONFIG_SYS_FSL_OCRAM_BASE, + MT_NORMAL); + for (i = CONFIG_SYS_IFC_BASE >> SECTION_SHIFT_L2; + i < (CONFIG_SYS_IFC_BASE + CONFIG_SYS_IFC_SIZE) + >> SECTION_SHIFT_L2; i++) { + section_l2 = i << SECTION_SHIFT_L2; + set_pgtable_section(level2_table, i, + section_l2, MT_NORMAL); + } + + el = current_el(); + set_ttbr_tcr_mair(el, (u64)level0_table, LSCH3_TCR, MEMORY_ATTRIBUTES); + set_sctlr(get_sctlr() | CR_M); +} + +/* + * This final tale looks similar to early table, but different in detail. + * These tables are in regular memory. Cache on IFC is disabled. One sub table + * is added to enable cache for QBMan. + */ +static inline void final_mmu_setup(void) +{ + int el; + u64 i, tbl_base, tbl_limit, section_base; + u64 section_l1t0, section_l1t1, section_l2; + u64 *level0_table = (u64 *)gd->arch.tlb_addr; + u64 *level1_table_0 = (u64 *)(gd->arch.tlb_addr + 0x1000); + u64 *level1_table_1 = (u64 *)(gd->arch.tlb_addr + 0x2000); + u64 *level2_table_0 = (u64 *)(gd->arch.tlb_addr + 0x3000); + u64 *level2_table_1 = (u64 *)(gd->arch.tlb_addr + 0x4000); + + + level0_table[0] = + (u64)level1_table_0 | PMD_TYPE_TABLE; + level0_table[1] = + (u64)level1_table_1 | PMD_TYPE_TABLE; + + /* + * set level 1 table 0 to cache_inhibit, covering 0 to 512GB + * set level 1 table 1 to cache enabled, covering 512GB to 1TB + * set level 2 table 0 to cache-inhibit, covering 0 to 1GB + */ + section_l1t0 = 0; + section_l1t1 = BLOCK_SIZE_L0; + section_l2 = 0; + for (i = 0; i < 512; i++) { + set_pgtable_section(level1_table_0, i, section_l1t0, + MT_DEVICE_NGNRNE); + set_pgtable_section(level1_table_1, i, section_l1t1, + MT_NORMAL); + set_pgtable_section(level2_table_0, i, section_l2, + MT_DEVICE_NGNRNE); + section_l1t0 += BLOCK_SIZE_L1; + section_l1t1 += BLOCK_SIZE_L1; + section_l2 += BLOCK_SIZE_L2; + } + + level1_table_0[0] = + (u64)level2_table_0 | PMD_TYPE_TABLE; + level1_table_0[2] = + 0x80000000 | PMD_SECT_AF | PMD_TYPE_SECT | + PMD_ATTRINDX(MT_NORMAL); + level1_table_0[3] = + 0xc0000000 | PMD_SECT_AF | PMD_TYPE_SECT | + PMD_ATTRINDX(MT_NORMAL); + + /* Rewrite table to enable cache */ + set_pgtable_section(level2_table_0, + CONFIG_SYS_FSL_OCRAM_BASE >> SECTION_SHIFT_L2, + CONFIG_SYS_FSL_OCRAM_BASE, + MT_NORMAL); + + /* + * Fill in other part of tables if cache is needed + * If finer granularity than 1GB is needed, sub table + * should be created. + */ + section_base = FINAL_QBMAN_CACHED_MEM & ~(BLOCK_SIZE_L1 - 1); + i = section_base >> SECTION_SHIFT_L1; + level1_table_0[i] = (u64)level2_table_1 | PMD_TYPE_TABLE; + section_l2 = section_base; + for (i = 0; i < 512; i++) { + set_pgtable_section(level2_table_1, i, section_l2, + MT_DEVICE_NGNRNE); + section_l2 += BLOCK_SIZE_L2; + } + tbl_base = FINAL_QBMAN_CACHED_MEM & (BLOCK_SIZE_L1 - 1); + tbl_limit = (FINAL_QBMAN_CACHED_MEM + FINAL_QBMAN_CACHED_SIZE) & + (BLOCK_SIZE_L1 - 1); + for (i = tbl_base >> SECTION_SHIFT_L2; + i < tbl_limit >> SECTION_SHIFT_L2; i++) { + section_l2 = section_base + (i << SECTION_SHIFT_L2); + set_pgtable_section(level2_table_1, i, + section_l2, MT_NORMAL); + } + + /* flush new MMU table */ + flush_dcache_range(gd->arch.tlb_addr, + gd->arch.tlb_addr + gd->arch.tlb_size); + + /* point TTBR to the new table */ + el = current_el(); + asm volatile("dsb sy"); + if (el == 1) { + asm volatile("msr ttbr0_el1, %0" + : : "r" ((u64)level0_table) : "memory"); + } else if (el == 2) { + asm volatile("msr ttbr0_el2, %0" + : : "r" ((u64)level0_table) : "memory"); + } else if (el == 3) { + asm volatile("msr ttbr0_el3, %0" + : : "r" ((u64)level0_table) : "memory"); + } else { + hang(); + } + asm volatile("isb"); + + /* + * MMU is already enabled, just need to invalidate TLB to load the + * new table. The new table is compatible with the current table, if + * MMU somehow walks through the new table before invalidation TLB, + * it still works. So we don't need to turn off MMU here. + */ +} + +int arch_cpu_init(void) +{ + icache_enable(); + __asm_invalidate_dcache_all(); + __asm_invalidate_tlb_all(); + early_mmu_setup(); + set_sctlr(get_sctlr() | CR_C); + return 0; +} + +/* + * flush_l3_cache + * Dickens L3 cache can be flushed by transitioning from FAM to SFONLY power + * state, by writing to HP-F P-state request register. + * Fixme: This function should moved to a common file if other SoCs also use + * the same Dickens. + */ +#define HNF0_PSTATE_REQ 0x04200010 +#define HNF1_PSTATE_REQ 0x04210010 +#define HNF2_PSTATE_REQ 0x04220010 +#define HNF3_PSTATE_REQ 0x04230010 +#define HNF4_PSTATE_REQ 0x04240010 +#define HNF5_PSTATE_REQ 0x04250010 +#define HNF6_PSTATE_REQ 0x04260010 +#define HNF7_PSTATE_REQ 0x04270010 +#define HNFPSTAT_MASK (0xFFFFFFFFFFFFFFFC) +#define HNFPSTAT_FAM 0x3 +#define HNFPSTAT_SFONLY 0x01 + +static void hnf_pstate_req(u64 *ptr, u64 state) +{ + int timeout = 1000; + out_le64(ptr, (in_le64(ptr) & HNFPSTAT_MASK) | (state & 0x3)); + ptr++; + /* checking if the transition is completed */ + while (timeout > 0) { + if (((in_le64(ptr) & 0x0c) >> 2) == (state & 0x3)) + break; + udelay(100); + timeout--; + } +} + +void flush_l3_cache(void) +{ + hnf_pstate_req((u64 *)HNF0_PSTATE_REQ, HNFPSTAT_SFONLY); + hnf_pstate_req((u64 *)HNF1_PSTATE_REQ, HNFPSTAT_SFONLY); + hnf_pstate_req((u64 *)HNF2_PSTATE_REQ, HNFPSTAT_SFONLY); + hnf_pstate_req((u64 *)HNF3_PSTATE_REQ, HNFPSTAT_SFONLY); + hnf_pstate_req((u64 *)HNF4_PSTATE_REQ, HNFPSTAT_SFONLY); + hnf_pstate_req((u64 *)HNF5_PSTATE_REQ, HNFPSTAT_SFONLY); + hnf_pstate_req((u64 *)HNF6_PSTATE_REQ, HNFPSTAT_SFONLY); + hnf_pstate_req((u64 *)HNF7_PSTATE_REQ, HNFPSTAT_SFONLY); + hnf_pstate_req((u64 *)HNF0_PSTATE_REQ, HNFPSTAT_FAM); + hnf_pstate_req((u64 *)HNF1_PSTATE_REQ, HNFPSTAT_FAM); + hnf_pstate_req((u64 *)HNF2_PSTATE_REQ, HNFPSTAT_FAM); + hnf_pstate_req((u64 *)HNF3_PSTATE_REQ, HNFPSTAT_FAM); + hnf_pstate_req((u64 *)HNF4_PSTATE_REQ, HNFPSTAT_FAM); + hnf_pstate_req((u64 *)HNF5_PSTATE_REQ, HNFPSTAT_FAM); + hnf_pstate_req((u64 *)HNF6_PSTATE_REQ, HNFPSTAT_FAM); + hnf_pstate_req((u64 *)HNF7_PSTATE_REQ, HNFPSTAT_FAM); +} + +/* + * This function is called from lib/board.c. + * It recreates MMU table in main memory. MMU and d-cache are enabled earlier. + * There is no need to disable d-cache for this operation. + */ +void enable_caches(void) +{ + final_mmu_setup(); + __asm_invalidate_tlb_all(); +} +#endif + +static inline u32 initiator_type(u32 cluster, int init_id) +{ + struct ccsr_gur *gur = (void *)(CONFIG_SYS_FSL_GUTS_ADDR); + u32 idx = (cluster >> (init_id * 8)) & TP_CLUSTER_INIT_MASK; + u32 type = in_le32(&gur->tp_ityp[idx]); + + if (type & TP_ITYP_AV) + return type; + + return 0; +} + +u32 cpu_mask(void) +{ + struct ccsr_gur __iomem *gur = (void *)(CONFIG_SYS_FSL_GUTS_ADDR); + int i = 0, count = 0; + u32 cluster, type, mask = 0; + + do { + int j; + cluster = in_le32(&gur->tp_cluster[i].lower); + for (j = 0; j < TP_INIT_PER_CLUSTER; j++) { + type = initiator_type(cluster, j); + if (type) { + if (TP_ITYP_TYPE(type) == TP_ITYP_TYPE_ARM) + mask |= 1 << count; + count++; + } + } + i++; + } while ((cluster & TP_CLUSTER_EOC) != TP_CLUSTER_EOC); + + return mask; +} + +/* + * Return the number of cores on this SOC. + */ +int cpu_numcores(void) +{ + return hweight32(cpu_mask()); +} + +int fsl_qoriq_core_to_cluster(unsigned int core) +{ + struct ccsr_gur __iomem *gur = + (void __iomem *)(CONFIG_SYS_FSL_GUTS_ADDR); + int i = 0, count = 0; + u32 cluster; + + do { + int j; + cluster = in_le32(&gur->tp_cluster[i].lower); + for (j = 0; j < TP_INIT_PER_CLUSTER; j++) { + if (initiator_type(cluster, j)) { + if (count == core) + return i; + count++; + } + } + i++; + } while ((cluster & TP_CLUSTER_EOC) != TP_CLUSTER_EOC); + + return -1; /* cannot identify the cluster */ +} + +u32 fsl_qoriq_core_to_type(unsigned int core) +{ + struct ccsr_gur __iomem *gur = + (void __iomem *)(CONFIG_SYS_FSL_GUTS_ADDR); + int i = 0, count = 0; + u32 cluster, type; + + do { + int j; + cluster = in_le32(&gur->tp_cluster[i].lower); + for (j = 0; j < TP_INIT_PER_CLUSTER; j++) { + type = initiator_type(cluster, j); + if (type) { + if (count == core) + return type; + count++; + } + } + i++; + } while ((cluster & TP_CLUSTER_EOC) != TP_CLUSTER_EOC); + + return -1; /* cannot identify the cluster */ +} + +#ifdef CONFIG_DISPLAY_CPUINFO +int print_cpuinfo(void) +{ + struct sys_info sysinfo; + char buf[32]; + unsigned int i, core; + u32 type; + + get_sys_info(&sysinfo); + puts("Clock Configuration:"); + for_each_cpu(i, core, cpu_numcores(), cpu_mask()) { + if (!(i % 3)) + puts("\n "); + type = TP_ITYP_VER(fsl_qoriq_core_to_type(core)); + printf("CPU%d(%s):%-4s MHz ", core, + type == TY_ITYP_VER_A7 ? "A7 " : + (type == TY_ITYP_VER_A53 ? "A53" : + (type == TY_ITYP_VER_A57 ? "A57" : " ")), + strmhz(buf, sysinfo.freq_processor[core])); + } + printf("\n Bus: %-4s MHz ", + strmhz(buf, sysinfo.freq_systembus)); + printf("DDR: %-4s MHz", strmhz(buf, sysinfo.freq_ddrbus)); + puts("\n"); + + return 0; +} +#endif -- cgit v1.2.1