/* SPDX-License-Identifier: GPL-2.0-only */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include /* This is a lot more paranoid now, since Linux can NOT handle * being told there is a CPU when none exists. So any errors * will return 0, meaning no CPU. * * We actually handling that case by noting which cpus startup * and not telling anyone about the ones that don't. */ /* Start-UP IPI vector must be 4kB aligned and below 1MB. */ #define AP_SIPI_VECTOR 0x1000 static char *lowmem_backup; static char *lowmem_backup_ptr; static int lowmem_backup_size; static inline void setup_secondary_gdt(void) { u16 *gdt_limit; #ifdef __x86_64__ u64 *gdt_base; #else u32 *gdt_base; #endif gdt_limit = (void *)&_secondary_gdt_addr; gdt_base = (void *)&gdt_limit[1]; *gdt_limit = (uintptr_t)&gdt_end - (uintptr_t)&gdt - 1; *gdt_base = (uintptr_t)&gdt; } static void copy_secondary_start_to_lowest_1M(void) { unsigned long code_size; /* Fill in secondary_start's local gdt. */ setup_secondary_gdt(); code_size = (unsigned long)_secondary_start_end - (unsigned long)_secondary_start; if (acpi_is_wakeup_s3()) { /* need to save it for RAM resume */ lowmem_backup_size = code_size; lowmem_backup = malloc(code_size); lowmem_backup_ptr = (char *)AP_SIPI_VECTOR; if (lowmem_backup == NULL) die("Out of backup memory\n"); memcpy(lowmem_backup, lowmem_backup_ptr, lowmem_backup_size); } /* copy the _secondary_start to the RAM below 1M*/ memcpy((unsigned char *)AP_SIPI_VECTOR, (unsigned char *)_secondary_start, code_size); printk(BIOS_DEBUG, "start_eip=0x%08lx, code_size=0x%08lx\n", (unsigned long int)AP_SIPI_VECTOR, code_size); } static void recover_lowest_1M(void) { if (acpi_is_wakeup_s3()) memcpy(lowmem_backup_ptr, lowmem_backup, lowmem_backup_size); } static int lapic_start_cpu(unsigned long apicid) { int timeout; unsigned long send_status, accept_status; int j, maxlvt; /* * Starting actual IPI sequence... */ printk(BIOS_SPEW, "Asserting INIT.\n"); /* * Turn INIT on target chip */ lapic_write_around(LAPIC_ICR2, SET_LAPIC_DEST_FIELD(apicid)); /* * Send IPI */ lapic_write_around(LAPIC_ICR, LAPIC_INT_LEVELTRIG | LAPIC_INT_ASSERT | LAPIC_DM_INIT); printk(BIOS_SPEW, "Waiting for send to finish...\n"); timeout = 0; do { printk(BIOS_SPEW, "+"); udelay(100); send_status = lapic_read(LAPIC_ICR) & LAPIC_ICR_BUSY; } while (send_status && (timeout++ < 1000)); if (timeout >= 1000) { printk(BIOS_ERR, "CPU %ld: First APIC write timed out. " "Disabling\n", apicid); // too bad. printk(BIOS_ERR, "ESR is 0x%lx\n", lapic_read(LAPIC_ESR)); if (lapic_read(LAPIC_ESR)) { printk(BIOS_ERR, "Try to reset ESR\n"); lapic_write_around(LAPIC_ESR, 0); printk(BIOS_ERR, "ESR is 0x%lx\n", lapic_read(LAPIC_ESR)); } return 0; } mdelay(10); printk(BIOS_SPEW, "Deasserting INIT.\n"); /* Target chip */ lapic_write_around(LAPIC_ICR2, SET_LAPIC_DEST_FIELD(apicid)); /* Send IPI */ lapic_write_around(LAPIC_ICR, LAPIC_INT_LEVELTRIG | LAPIC_DM_INIT); printk(BIOS_SPEW, "Waiting for send to finish...\n"); timeout = 0; do { printk(BIOS_SPEW, "+"); udelay(100); send_status = lapic_read(LAPIC_ICR) & LAPIC_ICR_BUSY; } while (send_status && (timeout++ < 1000)); if (timeout >= 1000) { printk(BIOS_ERR, "CPU %ld: Second APIC write timed out. " "Disabling\n", apicid); // too bad. return 0; } /* * Run STARTUP IPI loop. */ printk(BIOS_SPEW, "#startup loops: %d.\n", CONFIG_NUM_IPI_STARTS); maxlvt = 4; for (j = 1; j <= CONFIG_NUM_IPI_STARTS; j++) { printk(BIOS_SPEW, "Sending STARTUP #%d to %lu.\n", j, apicid); lapic_read_around(LAPIC_SPIV); lapic_write(LAPIC_ESR, 0); lapic_read(LAPIC_ESR); printk(BIOS_SPEW, "After apic_write.\n"); /* * STARTUP IPI */ /* Target chip */ lapic_write_around(LAPIC_ICR2, SET_LAPIC_DEST_FIELD(apicid)); /* Boot on the stack */ /* Kick the second */ lapic_write_around(LAPIC_ICR, LAPIC_DM_STARTUP | (AP_SIPI_VECTOR >> 12)); /* * Give the other CPU some time to accept the IPI. */ udelay(300); printk(BIOS_SPEW, "Startup point 1.\n"); printk(BIOS_SPEW, "Waiting for send to finish...\n"); timeout = 0; do { printk(BIOS_SPEW, "+"); udelay(100); send_status = lapic_read(LAPIC_ICR) & LAPIC_ICR_BUSY; } while (send_status && (timeout++ < 1000)); /* * Give the other CPU some time to accept the IPI. */ udelay(200); /* * Due to the Pentium erratum 3AP. */ if (maxlvt > 3) { lapic_read_around(LAPIC_SPIV); lapic_write(LAPIC_ESR, 0); } accept_status = (lapic_read(LAPIC_ESR) & 0xEF); if (send_status || accept_status) break; } printk(BIOS_SPEW, "After Startup.\n"); if (send_status) printk(BIOS_WARNING, "APIC never delivered???\n"); if (accept_status) printk(BIOS_WARNING, "APIC delivery error (%lx).\n", accept_status); if (send_status || accept_status) return 0; return 1; } /* Number of cpus that are currently running in coreboot */ static atomic_t active_cpus = ATOMIC_INIT(1); /* start_cpu_lock covers last_cpu_index and secondary_stack. * Only starting one CPU at a time let's me remove the logic * for select the stack from assembly language. * * In addition communicating by variables to the CPU I * am starting allows me to verify it has started before * start_cpu returns. */ DECLARE_SPIN_LOCK(start_cpu_lock); static unsigned int last_cpu_index = 0; static void *stacks[CONFIG_MAX_CPUS]; volatile unsigned long secondary_stack; volatile unsigned int secondary_cpu_index; int start_cpu(struct device *cpu) { struct cpu_info *info; uintptr_t stack_top; uintptr_t stack_base; unsigned long apicid; unsigned int index; unsigned long count; int result; spin_lock(&start_cpu_lock); /* Get the CPU's apicid */ apicid = cpu->path.apic.apic_id; /* Get an index for the new processor */ index = ++last_cpu_index; /* Find boundaries of the new processor's stack */ stack_top = ALIGN_DOWN((uintptr_t)_estack, CONFIG_STACK_SIZE); stack_top -= (CONFIG_STACK_SIZE*index); stack_base = stack_top - CONFIG_STACK_SIZE; stack_top -= sizeof(struct cpu_info); printk(BIOS_SPEW, "CPU%d: stack_base %p, stack_top %p\n", index, (void *)stack_base, (void *)stack_top); stacks[index] = (void *)stack_base; /* Record the index and which CPU structure we are using */ info = (struct cpu_info *)stack_top; info->index = index; info->cpu = cpu; cpu_add_map_entry(info->index); thread_init_cpu_info_non_bsp(info); /* Advertise the new stack and index to start_cpu */ secondary_stack = stack_top; secondary_cpu_index = index; /* Until the CPU starts up report the CPU is not enabled */ cpu->enabled = 0; cpu->initialized = 0; /* Start the CPU */ result = lapic_start_cpu(apicid); if (result) { result = 0; /* Wait 1s or until the new CPU calls in */ for (count = 0; count < 100000; count++) { if (secondary_stack == 0) { result = 1; break; } udelay(10); } } secondary_stack = 0; spin_unlock(&start_cpu_lock); return result; } #if CONFIG(AP_IN_SIPI_WAIT) /** * Sending INIT IPI to self is equivalent of asserting #INIT with a bit of * delay. * An undefined number of instruction cycles will complete. All global locks * must be released before INIT IPI and no printk is allowed after this. * De-asserting INIT IPI is a no-op on later Intel CPUs. * * If you set DEBUG_HALT_SELF to 1, printk's after INIT IPI are enabled * but running thread may halt without releasing the lock and effectively * deadlock other CPUs. */ #define DEBUG_HALT_SELF 0 /** * Normally this function is defined in lapic.h as an always inline function * that just keeps the CPU in a hlt() loop. This does not work on all CPUs. * I think all hyperthreading CPUs might need this version, but I could only * verify this on the Intel Core Duo */ void stop_this_cpu(void) { int timeout; unsigned long send_status; unsigned long id; id = lapicid(); printk(BIOS_DEBUG, "CPU %ld going down...\n", id); /* send an LAPIC INIT to myself */ lapic_write_around(LAPIC_ICR2, SET_LAPIC_DEST_FIELD(id)); lapic_write_around(LAPIC_ICR, LAPIC_INT_LEVELTRIG | LAPIC_INT_ASSERT | LAPIC_DM_INIT); /* wait for the ipi send to finish */ #if DEBUG_HALT_SELF printk(BIOS_SPEW, "Waiting for send to finish...\n"); #endif timeout = 0; do { #if DEBUG_HALT_SELF printk(BIOS_SPEW, "+"); #endif udelay(100); send_status = lapic_read(LAPIC_ICR) & LAPIC_ICR_BUSY; } while (send_status && (timeout++ < 1000)); if (timeout >= 1000) { #if DEBUG_HALT_SELF printk(BIOS_ERR, "timed out\n"); #endif } mdelay(10); #if DEBUG_HALT_SELF printk(BIOS_SPEW, "Deasserting INIT.\n"); #endif /* Deassert the LAPIC INIT */ lapic_write_around(LAPIC_ICR2, SET_LAPIC_DEST_FIELD(id)); lapic_write_around(LAPIC_ICR, LAPIC_INT_LEVELTRIG | LAPIC_DM_INIT); #if DEBUG_HALT_SELF printk(BIOS_SPEW, "Waiting for send to finish...\n"); #endif timeout = 0; do { #if DEBUG_HALT_SELF printk(BIOS_SPEW, "+"); #endif udelay(100); send_status = lapic_read(LAPIC_ICR) & LAPIC_ICR_BUSY; } while (send_status && (timeout++ < 1000)); if (timeout >= 1000) { #if DEBUG_HALT_SELF printk(BIOS_ERR, "timed out\n"); #endif } halt(); } #endif /* C entry point of secondary cpus */ asmlinkage void secondary_cpu_init(unsigned int index) { atomic_inc(&active_cpus); if (!CONFIG(PARALLEL_CPU_INIT)) spin_lock(&start_cpu_lock); #ifdef __SSE3__ /* * Seems that CR4 was cleared when AP start via lapic_start_cpu() * Turn on CR4.OSFXSR and CR4.OSXMMEXCPT when SSE options enabled */ CRx_TYPE cr4_val; cr4_val = read_cr4(); cr4_val |= (CR4_OSFXSR | CR4_OSXMMEXCPT); write_cr4(cr4_val); #endif cpu_initialize(index); if (!CONFIG(PARALLEL_CPU_INIT)) spin_unlock(&start_cpu_lock); atomic_dec(&active_cpus); stop_this_cpu(); } static void start_other_cpus(struct bus *cpu_bus, struct device *bsp_cpu) { struct device *cpu; /* Loop through the cpus once getting them started */ for (cpu = cpu_bus->children; cpu; cpu = cpu->sibling) { if (cpu->path.type != DEVICE_PATH_APIC) continue; if (CONFIG(PARALLEL_CPU_INIT) && (cpu == bsp_cpu)) continue; if (!cpu->enabled) continue; if (cpu->initialized) continue; if (!start_cpu(cpu)) /* Record the error in cpu? */ printk(BIOS_ERR, "CPU 0x%02x would not start!\n", cpu->path.apic.apic_id); if (!CONFIG(PARALLEL_CPU_INIT)) udelay(10); } } static void smm_other_cpus(struct bus *cpu_bus, struct device *bsp_cpu) { struct device *cpu; int pre_count = atomic_read(&active_cpus); /* Loop through the cpus once to let them run through SMM relocator */ for (cpu = cpu_bus->children; cpu; cpu = cpu->sibling) { if (cpu->path.type != DEVICE_PATH_APIC) continue; printk(BIOS_ERR, "considering CPU 0x%02x for SMM init\n", cpu->path.apic.apic_id); if (cpu == bsp_cpu) continue; if (!cpu->enabled) continue; if (!start_cpu(cpu)) /* Record the error in cpu? */ printk(BIOS_ERR, "CPU 0x%02x would not start!\n", cpu->path.apic.apic_id); /* FIXME: endless loop */ while (atomic_read(&active_cpus) != pre_count) ; } } static void wait_other_cpus_stop(struct bus *cpu_bus) { struct device *cpu; int old_active_count, active_count; long loopcount = 0; int i; /* Now loop until the other cpus have finished initializing */ old_active_count = 1; active_count = atomic_read(&active_cpus); while (active_count > 1) { if (active_count != old_active_count) { printk(BIOS_INFO, "Waiting for %d CPUS to stop\n", active_count - 1); old_active_count = active_count; } udelay(10); active_count = atomic_read(&active_cpus); loopcount++; } for (cpu = cpu_bus->children; cpu; cpu = cpu->sibling) { if (cpu->path.type != DEVICE_PATH_APIC) continue; if (cpu->path.apic.apic_id == SPEEDSTEP_APIC_MAGIC) continue; if (!cpu->initialized) printk(BIOS_ERR, "CPU 0x%02x did not initialize!\n", cpu->path.apic.apic_id); } printk(BIOS_DEBUG, "All AP CPUs stopped (%ld loops)\n", loopcount); checkstack(_estack, 0); for (i = 1; i < CONFIG_MAX_CPUS && i <= last_cpu_index; i++) checkstack((void *)stacks[i] + CONFIG_STACK_SIZE, i); } void initialize_cpus(struct bus *cpu_bus) { struct device_path cpu_path; struct cpu_info *info; /* Find the info struct for this CPU */ info = cpu_info(); if (need_lapic_init()) { /* Ensure the local APIC is enabled */ enable_lapic(); /* Get the device path of the boot CPU */ cpu_path.type = DEVICE_PATH_APIC; cpu_path.apic.apic_id = lapicid(); } else { /* Get the device path of the boot CPU */ cpu_path.type = DEVICE_PATH_CPU; cpu_path.cpu.id = 0; } /* Find the device structure for the boot CPU */ info->cpu = alloc_find_dev(cpu_bus, &cpu_path); cpu_add_map_entry(info->index); // why here? In case some day we can start core1 in amd_sibling_init if (is_smp_boot()) copy_secondary_start_to_lowest_1M(); if (!CONFIG(SERIALIZED_SMM_INITIALIZATION)) smm_init(); /* start all aps at first, so we can init ECC all together */ if (is_smp_boot() && CONFIG(PARALLEL_CPU_INIT)) start_other_cpus(cpu_bus, info->cpu); /* Initialize the bootstrap processor */ cpu_initialize(0); if (is_smp_boot() && !CONFIG(PARALLEL_CPU_INIT)) start_other_cpus(cpu_bus, info->cpu); /* Now wait the rest of the cpus stop*/ if (is_smp_boot()) wait_other_cpus_stop(cpu_bus); if (CONFIG(SERIALIZED_SMM_INITIALIZATION)) { /* At this point, all APs are sleeping: * smm_init() will queue a pending SMI on all cpus * and smm_other_cpus() will start them one by one */ smm_init(); if (is_smp_boot()) { last_cpu_index = 0; smm_other_cpus(cpu_bus, info->cpu); } } smm_init_completion(); if (is_smp_boot()) recover_lowest_1M(); }