From decad308655edf61094d5d552ab93f9fb2a6d535 Mon Sep 17 00:00:00 2001 From: Anton Fedotov Date: Wed, 14 Apr 2010 07:36:08 +0200 Subject: [PATCH] cortex-a8: more MMU support MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit + virt2phys() can now convert virtual address to real + read_memory() and write_memory() are renamed to read_phys_memory() and write_phys_memory() + new read_memory() and write_memory() try to resolve real address if mmu is enambled than perform real address reading/writing + if address is bellow 0xc000000 than TTB0 is used for page table dereference, if above - than TTB1. Linux style of user/kernel address separation + if above fails (i.e address is unspecified) than mode is checked whether it is Supervisor (than TTB1) or User (than TTB0) - Software breakpoints doesn't work. You should invoke "gdb_breakpoint_override hard" before you start debugging + cortex_a8_mmu(), cortex_a8_enable_mmu_caches(), cortex_a8_disable_mmu_caches() are implemented Signed-off-by: Øyvind Harboe --- src/target/cortex_a8.c | 408 +++++++++++++++++++++++++++++++---------- src/target/cortex_a8.h | 3 + 2 files changed, 317 insertions(+), 94 deletions(-) diff --git a/src/target/cortex_a8.c b/src/target/cortex_a8.c index a5480272e..aa0e030cc 100644 --- a/src/target/cortex_a8.c +++ b/src/target/cortex_a8.c @@ -51,6 +51,16 @@ static int cortex_a8_dap_read_coreregister_u32(struct target *target, uint32_t *value, int regnum); static int cortex_a8_dap_write_coreregister_u32(struct target *target, uint32_t value, int regnum); +static int cortex_a8_mmu(struct target *target, int *enabled); +static int cortex_a8_virt2phys(struct target *target, + uint32_t virt, uint32_t *phys); +static void cortex_a8_disable_mmu_caches(struct target *target, int mmu, + int d_u_cache, int i_cache); +static void cortex_a8_enable_mmu_caches(struct target *target, int mmu, + int d_u_cache, int i_cache); +static uint32_t cortex_a8_get_ttb(struct target *target); + + /* * FIXME do topology discovery using the ROM; don't * assume this is an OMAP3. Also, allow for multiple ARMv7-A @@ -1269,111 +1279,158 @@ static int cortex_a8_deassert_reset(struct target *target) * ap number for every access. */ +static int cortex_a8_read_phys_memory(struct target *target, + uint32_t address, uint32_t size, + uint32_t count, uint8_t *buffer) +{ + struct armv7a_common *armv7a = target_to_armv7a(target); + struct adiv5_dap *swjdp = &armv7a->dap; + int retval = ERROR_INVALID_ARGUMENTS; + + /* cortex_a8 handles unaligned memory access */ + +// ??? dap_ap_select(swjdp, swjdp_memoryap); + LOG_DEBUG("Reading memory at real address 0x%x; size %d; count %d", address, size, count); + if (count && buffer) { + switch (size) { + case 4: + retval = mem_ap_read_buf_u32(swjdp, buffer, 4 * count, address); + break; + case 2: + retval = mem_ap_read_buf_u16(swjdp, buffer, 2 * count, address); + break; + case 1: + retval = mem_ap_read_buf_u8(swjdp, buffer, count, address); + break; + } + } + + return retval; +} + static int cortex_a8_read_memory(struct target *target, uint32_t address, uint32_t size, uint32_t count, uint8_t *buffer) { - struct armv7a_common *armv7a = target_to_armv7a(target); - struct adiv5_dap *swjdp = &armv7a->dap; - int retval = ERROR_INVALID_ARGUMENTS; + int enabled = 0; + uint32_t virt, phys; /* cortex_a8 handles unaligned memory access */ // ??? dap_ap_select(swjdp, swjdp_memoryap); + LOG_DEBUG("Reading memory at address 0x%x; size %d; count %d", address, size, count); + cortex_a8_mmu(target, &enabled); + if(enabled) + { + virt = address; + cortex_a8_virt2phys(target, virt, &phys); + LOG_DEBUG("Reading at virtual address. Translating v:0x%x to r:0x%x", virt, phys); + address = phys; + } - if (count && buffer) { - switch (size) { - case 4: - retval = mem_ap_read_buf_u32(swjdp, buffer, 4 * count, address); - break; - case 2: - retval = mem_ap_read_buf_u16(swjdp, buffer, 2 * count, address); - break; - case 1: - retval = mem_ap_read_buf_u8(swjdp, buffer, count, address); - break; - } - } + return cortex_a8_read_phys_memory(target, address, size, count, buffer); +} - return retval; +static int cortex_a8_write_phys_memory(struct target *target, + uint32_t address, uint32_t size, + uint32_t count, uint8_t *buffer) +{ + struct armv7a_common *armv7a = target_to_armv7a(target); + struct adiv5_dap *swjdp = &armv7a->dap; + int retval = ERROR_INVALID_ARGUMENTS; + +// ??? dap_ap_select(swjdp, swjdp_memoryap); + + LOG_DEBUG("Writing memory to real address 0x%x; size %d; count %d", address, size, count); + if (count && buffer) { + switch (size) { + case 4: + retval = mem_ap_write_buf_u32(swjdp, buffer, 4 * count, address); + break; + case 2: + retval = mem_ap_write_buf_u16(swjdp, buffer, 2 * count, address); + break; + case 1: + retval = mem_ap_write_buf_u8(swjdp, buffer, count, address); + break; + } + } + + /* REVISIT this op is generic ARMv7-A/R stuff */ + if (retval == ERROR_OK && target->state == TARGET_HALTED) + { + struct arm_dpm *dpm = armv7a->armv4_5_common.dpm; + + retval = dpm->prepare(dpm); + if (retval != ERROR_OK) + return retval; + + /* The Cache handling will NOT work with MMU active, the + * wrong addresses will be invalidated! + * + * For both ICache and DCache, walk all cache lines in the + * address range. Cortex-A8 has fixed 64 byte line length. + * + * REVISIT per ARMv7, these may trigger watchpoints ... + */ + + /* invalidate I-Cache */ + if (armv7a->armv4_5_mmu.armv4_5_cache.i_cache_enabled) + { + /* ICIMVAU - Invalidate Cache single entry + * with MVA to PoU + * MCR p15, 0, r0, c7, c5, 1 + */ + for (uint32_t cacheline = address; + cacheline < address + size * count; + cacheline += 64) { + retval = dpm->instr_write_data_r0(dpm, + ARMV4_5_MCR(15, 0, 0, 7, 5, 1), + cacheline); + } + } + + /* invalidate D-Cache */ + if (armv7a->armv4_5_mmu.armv4_5_cache.d_u_cache_enabled) + { + /* DCIMVAC - Invalidate data Cache line + * with MVA to PoC + * MCR p15, 0, r0, c7, c6, 1 + */ + for (uint32_t cacheline = address; + cacheline < address + size * count; + cacheline += 64) { + retval = dpm->instr_write_data_r0(dpm, + ARMV4_5_MCR(15, 0, 0, 7, 6, 1), + cacheline); + } + } + + /* (void) */ dpm->finish(dpm); + } + + return retval; } static int cortex_a8_write_memory(struct target *target, uint32_t address, - uint32_t size, uint32_t count, uint8_t *buffer) + uint32_t size, uint32_t count, uint8_t *buffer) { - struct armv7a_common *armv7a = target_to_armv7a(target); - struct adiv5_dap *swjdp = &armv7a->dap; - int retval = ERROR_INVALID_ARGUMENTS; + int enabled = 0; + uint32_t virt, phys; -// ??? dap_ap_select(swjdp, swjdp_memoryap); +// ??? dap_ap_select(swjdp, swjdp_memoryap); - if (count && buffer) { - switch (size) { - case 4: - retval = mem_ap_write_buf_u32(swjdp, buffer, 4 * count, address); - break; - case 2: - retval = mem_ap_write_buf_u16(swjdp, buffer, 2 * count, address); - break; - case 1: - retval = mem_ap_write_buf_u8(swjdp, buffer, count, address); - break; - } - } + LOG_DEBUG("Writing memory to address 0x%x; size %d; count %d", address, size, count); + cortex_a8_mmu(target, &enabled); + if(enabled) + { + virt = address; + cortex_a8_virt2phys(target, virt, &phys); + LOG_DEBUG("Writing to virtual address. Translating v:0x%x to r:0x%x", virt, phys); + address = phys; + } - /* REVISIT this op is generic ARMv7-A/R stuff */ - if (retval == ERROR_OK && target->state == TARGET_HALTED) - { - struct arm_dpm *dpm = armv7a->armv4_5_common.dpm; - - retval = dpm->prepare(dpm); - if (retval != ERROR_OK) - return retval; - - /* The Cache handling will NOT work with MMU active, the - * wrong addresses will be invalidated! - * - * For both ICache and DCache, walk all cache lines in the - * address range. Cortex-A8 has fixed 64 byte line length. - * - * REVISIT per ARMv7, these may trigger watchpoints ... - */ - - /* invalidate I-Cache */ - if (armv7a->armv4_5_mmu.armv4_5_cache.i_cache_enabled) - { - /* ICIMVAU - Invalidate Cache single entry - * with MVA to PoU - * MCR p15, 0, r0, c7, c5, 1 - */ - for (uint32_t cacheline = address; - cacheline < address + size * count; - cacheline += 64) { - retval = dpm->instr_write_data_r0(dpm, - ARMV4_5_MCR(15, 0, 0, 7, 5, 1), - cacheline); - } - } - - /* invalidate D-Cache */ - if (armv7a->armv4_5_mmu.armv4_5_cache.d_u_cache_enabled) - { - /* DCIMVAC - Invalidate data Cache line - * with MVA to PoC - * MCR p15, 0, r0, c7, c6, 1 - */ - for (uint32_t cacheline = address; - cacheline < address + size * count; - cacheline += 64) { - retval = dpm->instr_write_data_r0(dpm, - ARMV4_5_MCR(15, 0, 0, 7, 6, 1), - cacheline); - } - } - - /* (void) */ dpm->finish(dpm); - } - - return retval; + return cortex_a8_write_phys_memory(target, address, size, + count, buffer); } static int cortex_a8_bulk_write_memory(struct target *target, uint32_t address, @@ -1580,6 +1637,9 @@ static int cortex_a8_init_arch_info(struct target *target, cortex_a8->fast_reg_read = 0; + /* Set default value */ + cortex_a8->current_address_mode = ARM_MODE_ANY; + /* register arch-specific functions */ armv7a->examine_debug_reason = NULL; @@ -1587,11 +1647,11 @@ static int cortex_a8_init_arch_info(struct target *target, armv7a->pre_restore_context = NULL; armv7a->armv4_5_mmu.armv4_5_cache.ctype = -1; -// armv7a->armv4_5_mmu.get_ttb = armv7a_get_ttb; - armv7a->armv4_5_mmu.read_memory = cortex_a8_read_memory; - armv7a->armv4_5_mmu.write_memory = cortex_a8_write_memory; -// armv7a->armv4_5_mmu.disable_mmu_caches = armv7a_disable_mmu_caches; -// armv7a->armv4_5_mmu.enable_mmu_caches = armv7a_enable_mmu_caches; + armv7a->armv4_5_mmu.get_ttb = cortex_a8_get_ttb; + armv7a->armv4_5_mmu.read_memory = cortex_a8_read_phys_memory; + armv7a->armv4_5_mmu.write_memory = cortex_a8_write_phys_memory; + armv7a->armv4_5_mmu.disable_mmu_caches = cortex_a8_disable_mmu_caches; + armv7a->armv4_5_mmu.enable_mmu_caches = cortex_a8_enable_mmu_caches; armv7a->armv4_5_mmu.has_tiny_pages = 1; armv7a->armv4_5_mmu.mmu_enabled = 0; @@ -1616,6 +1676,160 @@ static int cortex_a8_target_create(struct target *target, Jim_Interp *interp) return ERROR_OK; } +static uint32_t cortex_a8_get_ttb(struct target *target) +{ + struct cortex_a8_common *cortex_a8 = target_to_cortex_a8(target); + struct armv7a_common *armv7a = &cortex_a8->armv7a_common; + uint32_t ttb = 0, retval = ERROR_OK; + + /* current_address_mode is set inside cortex_a8_virt2phys() + where we can determine if address belongs to user or kernel */ + if(cortex_a8->current_address_mode == ARM_MODE_SVC) + { + /* MRC p15,0,,c1,c0,0 ; Read CP15 System Control Register */ + retval = armv7a->armv4_5_common.mrc(target, 15, + 0, 1, /* op1, op2 */ + 2, 0, /* CRn, CRm */ + &ttb); + } + else if(cortex_a8->current_address_mode == ARM_MODE_USR) + { + /* MRC p15,0,,c1,c0,0 ; Read CP15 System Control Register */ + retval = armv7a->armv4_5_common.mrc(target, 15, + 0, 0, /* op1, op2 */ + 2, 0, /* CRn, CRm */ + &ttb); + } + /* we don't know whose address is: user or kernel + we assume that if we are in kernel mode then + address belongs to kernel else if in user mode + - to user */ + else if(armv7a->armv4_5_common.core_mode == ARM_MODE_SVC) + { + /* MRC p15,0,,c1,c0,0 ; Read CP15 System Control Register */ + retval = armv7a->armv4_5_common.mrc(target, 15, + 0, 1, /* op1, op2 */ + 2, 0, /* CRn, CRm */ + &ttb); + } + else if(armv7a->armv4_5_common.core_mode == ARM_MODE_USR) + { + /* MRC p15,0,,c1,c0,0 ; Read CP15 System Control Register */ + retval = armv7a->armv4_5_common.mrc(target, 15, + 0, 0, /* op1, op2 */ + 2, 0, /* CRn, CRm */ + &ttb); + } + /* finaly we don't know whose ttb to use: user or kernel */ + else + LOG_ERROR("Don't know how to get ttb for current mode!!!"); + + ttb &= 0xffffc000; + + return ttb; +} + +static void cortex_a8_disable_mmu_caches(struct target *target, int mmu, + int d_u_cache, int i_cache) +{ + struct cortex_a8_common *cortex_a8 = target_to_cortex_a8(target); + struct armv7a_common *armv7a = &cortex_a8->armv7a_common; + uint32_t cp15_control; + + /* read cp15 control register */ + armv7a->armv4_5_common.mrc(target, 15, + 0, 0, /* op1, op2 */ + 1, 0, /* CRn, CRm */ + &cp15_control); + + + if (mmu) + cp15_control &= ~0x1U; + + if (d_u_cache) + cp15_control &= ~0x4U; + + if (i_cache) + cp15_control &= ~0x1000U; + + armv7a->armv4_5_common.mcr(target, 15, + 0, 0, /* op1, op2 */ + 1, 0, /* CRn, CRm */ + cp15_control); +} + +static void cortex_a8_enable_mmu_caches(struct target *target, int mmu, + int d_u_cache, int i_cache) +{ + struct cortex_a8_common *cortex_a8 = target_to_cortex_a8(target); + struct armv7a_common *armv7a = &cortex_a8->armv7a_common; + uint32_t cp15_control; + + /* read cp15 control register */ + armv7a->armv4_5_common.mrc(target, 15, + 0, 0, /* op1, op2 */ + 1, 0, /* CRn, CRm */ + &cp15_control); + + if (mmu) + cp15_control |= 0x1U; + + if (d_u_cache) + cp15_control |= 0x4U; + + if (i_cache) + cp15_control |= 0x1000U; + + armv7a->armv4_5_common.mcr(target, 15, + 0, 0, /* op1, op2 */ + 1, 0, /* CRn, CRm */ + cp15_control); +} + + +static int cortex_a8_mmu(struct target *target, int *enabled) +{ + if (target->state != TARGET_HALTED) { + LOG_ERROR("%s: target not halted", __func__); + return ERROR_TARGET_INVALID; + } + + *enabled = target_to_cortex_a8(target)->armv7a_common.armv4_5_mmu.mmu_enabled; + return ERROR_OK; +} + +static int cortex_a8_virt2phys(struct target *target, + uint32_t virt, uint32_t *phys) +{ + int type; + uint32_t cb; + int domain; + uint32_t ap; + struct cortex_a8_common *cortex_a8 = target_to_cortex_a8(target); + // struct armv7a_common *armv7a = &cortex_a8->armv7a_common; + struct armv7a_common *armv7a = target_to_armv7a(target); + + /* We assume that virtual address is separated + between user and kernel in Linux style: + 0x00000000-0xbfffffff - User space + 0xc0000000-0xffffffff - Kernel space */ + if( virt < 0xc0000000 ) /* Linux user space */ + cortex_a8->current_address_mode = ARM_MODE_USR; + else /* Linux kernel */ + cortex_a8->current_address_mode = ARM_MODE_SVC; + uint32_t ret = armv4_5_mmu_translate_va(target, + &armv7a->armv4_5_mmu, virt, &type, &cb, &domain, &ap); + /* Reset the flag. We don't want someone else to use it by error */ + cortex_a8->current_address_mode = ARM_MODE_ANY; + + if (type == -1) + { + return ret; + } + *phys = ret; + return ERROR_OK; +} + COMMAND_HANDLER(cortex_a8_handle_cache_info_command) { struct target *target = get_current_target(CMD_CTX); @@ -1703,4 +1917,10 @@ struct target_type cortexa8_target = { .target_create = cortex_a8_target_create, .init_target = cortex_a8_init_target, .examine = cortex_a8_examine, + + .read_phys_memory = cortex_a8_read_phys_memory, + .write_phys_memory = cortex_a8_write_phys_memory, + .mmu = cortex_a8_mmu, + .virt2phys = cortex_a8_virt2phys, + }; diff --git a/src/target/cortex_a8.h b/src/target/cortex_a8.h index cc2e0090d..7e116e49a 100644 --- a/src/target/cortex_a8.h +++ b/src/target/cortex_a8.h @@ -72,6 +72,9 @@ struct cortex_a8_common /* Use cortex_a8_read_regs_through_mem for fast register reads */ int fast_reg_read; + /* Flag that helps to resolve what ttb to use: user or kernel */ + int current_address_mode; + struct armv7a_common armv7a_common; };