diff --git a/arch/powerpc/include/asm/fadump.h b/arch/powerpc/include/asm/fadump.h index a6774560afe3753dd0f4636124c1021305b7a391..493e72f64b35feac4a03fe7912aa7c04067e6fe6 100644 --- a/arch/powerpc/include/asm/fadump.h +++ b/arch/powerpc/include/asm/fadump.h @@ -70,39 +70,39 @@ #define CPU_UNKNOWN (~((u32)0)) /* Utility macros */ -#define SKIP_TO_NEXT_CPU(reg_entry) \ -({ \ - while (reg_entry->reg_id != REG_ID("CPUEND")) \ - reg_entry++; \ - reg_entry++; \ +#define SKIP_TO_NEXT_CPU(reg_entry) \ +({ \ + while (be64_to_cpu(reg_entry->reg_id) != REG_ID("CPUEND")) \ + reg_entry++; \ + reg_entry++; \ }) /* Kernel Dump section info */ struct fadump_section { - u32 request_flag; - u16 source_data_type; - u16 error_flags; - u64 source_address; - u64 source_len; - u64 bytes_dumped; - u64 destination_address; + __be32 request_flag; + __be16 source_data_type; + __be16 error_flags; + __be64 source_address; + __be64 source_len; + __be64 bytes_dumped; + __be64 destination_address; }; /* ibm,configure-kernel-dump header. */ struct fadump_section_header { - u32 dump_format_version; - u16 dump_num_sections; - u16 dump_status_flag; - u32 offset_first_dump_section; + __be32 dump_format_version; + __be16 dump_num_sections; + __be16 dump_status_flag; + __be32 offset_first_dump_section; /* Fields for disk dump option. */ - u32 dd_block_size; - u64 dd_block_offset; - u64 dd_num_blocks; - u32 dd_offset_disk_path; + __be32 dd_block_size; + __be64 dd_block_offset; + __be64 dd_num_blocks; + __be32 dd_offset_disk_path; /* Maximum time allowed to prevent an automatic dump-reboot. */ - u32 max_time_auto; + __be32 max_time_auto; }; /* @@ -174,15 +174,15 @@ static inline u64 str_to_u64(const char *str) /* Register save area header. */ struct fadump_reg_save_area_header { - u64 magic_number; - u32 version; - u32 num_cpu_offset; + __be64 magic_number; + __be32 version; + __be32 num_cpu_offset; }; /* Register entry. */ struct fadump_reg_entry { - u64 reg_id; - u64 reg_value; + __be64 reg_id; + __be64 reg_value; }; /* fadump crash info structure */ diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S index 5bbd1bc8c3b0a3f4265e263ce315311d8562f689..0905c8da90f1bb6e2aedb5e4cbfd7a7351df44a7 100644 --- a/arch/powerpc/kernel/entry_64.S +++ b/arch/powerpc/kernel/entry_64.S @@ -659,7 +659,13 @@ _GLOBAL(ret_from_except_lite) 3: #endif bl save_nvgprs + /* + * Use a non volatile GPR to save and restore our thread_info flags + * across the call to restore_interrupts. + */ + mr r30,r4 bl restore_interrupts + mr r4,r30 addi r3,r1,STACK_FRAME_OVERHEAD bl do_notify_resume b ret_from_except diff --git a/arch/powerpc/kernel/fadump.c b/arch/powerpc/kernel/fadump.c index 742694c1d85238fe380e4ceecd7443506bb8d0b7..26d091a1a54cf555627fd95bb135cf98eb3e52b8 100644 --- a/arch/powerpc/kernel/fadump.c +++ b/arch/powerpc/kernel/fadump.c @@ -58,7 +58,7 @@ int __init early_init_dt_scan_fw_dump(unsigned long node, const __be32 *sections; int i, num_sections; int size; - const int *token; + const __be32 *token; if (depth != 1 || strcmp(uname, "rtas") != 0) return 0; @@ -72,7 +72,7 @@ int __init early_init_dt_scan_fw_dump(unsigned long node, return 1; fw_dump.fadump_supported = 1; - fw_dump.ibm_configure_kernel_dump = *token; + fw_dump.ibm_configure_kernel_dump = be32_to_cpu(*token); /* * The 'ibm,kernel-dump' rtas node is present only if there is @@ -147,11 +147,11 @@ static unsigned long init_fadump_mem_struct(struct fadump_mem_struct *fdm, memset(fdm, 0, sizeof(struct fadump_mem_struct)); addr = addr & PAGE_MASK; - fdm->header.dump_format_version = 0x00000001; - fdm->header.dump_num_sections = 3; + fdm->header.dump_format_version = cpu_to_be32(0x00000001); + fdm->header.dump_num_sections = cpu_to_be16(3); fdm->header.dump_status_flag = 0; fdm->header.offset_first_dump_section = - (u32)offsetof(struct fadump_mem_struct, cpu_state_data); + cpu_to_be32((u32)offsetof(struct fadump_mem_struct, cpu_state_data)); /* * Fields for disk dump option. @@ -167,27 +167,27 @@ static unsigned long init_fadump_mem_struct(struct fadump_mem_struct *fdm, /* Kernel dump sections */ /* cpu state data section. */ - fdm->cpu_state_data.request_flag = FADUMP_REQUEST_FLAG; - fdm->cpu_state_data.source_data_type = FADUMP_CPU_STATE_DATA; + fdm->cpu_state_data.request_flag = cpu_to_be32(FADUMP_REQUEST_FLAG); + fdm->cpu_state_data.source_data_type = cpu_to_be16(FADUMP_CPU_STATE_DATA); fdm->cpu_state_data.source_address = 0; - fdm->cpu_state_data.source_len = fw_dump.cpu_state_data_size; - fdm->cpu_state_data.destination_address = addr; + fdm->cpu_state_data.source_len = cpu_to_be64(fw_dump.cpu_state_data_size); + fdm->cpu_state_data.destination_address = cpu_to_be64(addr); addr += fw_dump.cpu_state_data_size; /* hpte region section */ - fdm->hpte_region.request_flag = FADUMP_REQUEST_FLAG; - fdm->hpte_region.source_data_type = FADUMP_HPTE_REGION; + fdm->hpte_region.request_flag = cpu_to_be32(FADUMP_REQUEST_FLAG); + fdm->hpte_region.source_data_type = cpu_to_be16(FADUMP_HPTE_REGION); fdm->hpte_region.source_address = 0; - fdm->hpte_region.source_len = fw_dump.hpte_region_size; - fdm->hpte_region.destination_address = addr; + fdm->hpte_region.source_len = cpu_to_be64(fw_dump.hpte_region_size); + fdm->hpte_region.destination_address = cpu_to_be64(addr); addr += fw_dump.hpte_region_size; /* RMA region section */ - fdm->rmr_region.request_flag = FADUMP_REQUEST_FLAG; - fdm->rmr_region.source_data_type = FADUMP_REAL_MODE_REGION; - fdm->rmr_region.source_address = RMA_START; - fdm->rmr_region.source_len = fw_dump.boot_memory_size; - fdm->rmr_region.destination_address = addr; + fdm->rmr_region.request_flag = cpu_to_be32(FADUMP_REQUEST_FLAG); + fdm->rmr_region.source_data_type = cpu_to_be16(FADUMP_REAL_MODE_REGION); + fdm->rmr_region.source_address = cpu_to_be64(RMA_START); + fdm->rmr_region.source_len = cpu_to_be64(fw_dump.boot_memory_size); + fdm->rmr_region.destination_address = cpu_to_be64(addr); addr += fw_dump.boot_memory_size; return addr; @@ -272,7 +272,7 @@ int __init fadump_reserve_mem(void) * first kernel. */ if (fdm_active) - fw_dump.boot_memory_size = fdm_active->rmr_region.source_len; + fw_dump.boot_memory_size = be64_to_cpu(fdm_active->rmr_region.source_len); else fw_dump.boot_memory_size = fadump_calculate_reserve_size(); @@ -314,8 +314,8 @@ int __init fadump_reserve_mem(void) (unsigned long)(base >> 20)); fw_dump.fadumphdr_addr = - fdm_active->rmr_region.destination_address + - fdm_active->rmr_region.source_len; + be64_to_cpu(fdm_active->rmr_region.destination_address) + + be64_to_cpu(fdm_active->rmr_region.source_len); pr_debug("fadumphdr_addr = %p\n", (void *) fw_dump.fadumphdr_addr); } else { @@ -472,9 +472,9 @@ fadump_read_registers(struct fadump_reg_entry *reg_entry, struct pt_regs *regs) { memset(regs, 0, sizeof(struct pt_regs)); - while (reg_entry->reg_id != REG_ID("CPUEND")) { - fadump_set_regval(regs, reg_entry->reg_id, - reg_entry->reg_value); + while (be64_to_cpu(reg_entry->reg_id) != REG_ID("CPUEND")) { + fadump_set_regval(regs, be64_to_cpu(reg_entry->reg_id), + be64_to_cpu(reg_entry->reg_value)); reg_entry++; } reg_entry++; @@ -603,20 +603,20 @@ static int __init fadump_build_cpu_notes(const struct fadump_mem_struct *fdm) if (!fdm->cpu_state_data.bytes_dumped) return -EINVAL; - addr = fdm->cpu_state_data.destination_address; + addr = be64_to_cpu(fdm->cpu_state_data.destination_address); vaddr = __va(addr); reg_header = vaddr; - if (reg_header->magic_number != REGSAVE_AREA_MAGIC) { + if (be64_to_cpu(reg_header->magic_number) != REGSAVE_AREA_MAGIC) { printk(KERN_ERR "Unable to read register save area.\n"); return -ENOENT; } pr_debug("--------CPU State Data------------\n"); - pr_debug("Magic Number: %llx\n", reg_header->magic_number); - pr_debug("NumCpuOffset: %x\n", reg_header->num_cpu_offset); + pr_debug("Magic Number: %llx\n", be64_to_cpu(reg_header->magic_number)); + pr_debug("NumCpuOffset: %x\n", be32_to_cpu(reg_header->num_cpu_offset)); - vaddr += reg_header->num_cpu_offset; - num_cpus = *((u32 *)(vaddr)); + vaddr += be32_to_cpu(reg_header->num_cpu_offset); + num_cpus = be32_to_cpu(*((__be32 *)(vaddr))); pr_debug("NumCpus : %u\n", num_cpus); vaddr += sizeof(u32); reg_entry = (struct fadump_reg_entry *)vaddr; @@ -639,13 +639,13 @@ static int __init fadump_build_cpu_notes(const struct fadump_mem_struct *fdm) fdh = __va(fw_dump.fadumphdr_addr); for (i = 0; i < num_cpus; i++) { - if (reg_entry->reg_id != REG_ID("CPUSTRT")) { + if (be64_to_cpu(reg_entry->reg_id) != REG_ID("CPUSTRT")) { printk(KERN_ERR "Unable to read CPU state data\n"); rc = -ENOENT; goto error_out; } /* Lower 4 bytes of reg_value contains logical cpu id */ - cpu = reg_entry->reg_value & FADUMP_CPU_ID_MASK; + cpu = be64_to_cpu(reg_entry->reg_value) & FADUMP_CPU_ID_MASK; if (fdh && !cpumask_test_cpu(cpu, &fdh->cpu_online_mask)) { SKIP_TO_NEXT_CPU(reg_entry); continue; @@ -692,7 +692,7 @@ static int __init process_fadump(const struct fadump_mem_struct *fdm_active) return -EINVAL; /* Check if the dump data is valid. */ - if ((fdm_active->header.dump_status_flag == FADUMP_ERROR_FLAG) || + if ((be16_to_cpu(fdm_active->header.dump_status_flag) == FADUMP_ERROR_FLAG) || (fdm_active->cpu_state_data.error_flags != 0) || (fdm_active->rmr_region.error_flags != 0)) { printk(KERN_ERR "Dump taken by platform is not valid\n"); @@ -828,7 +828,7 @@ static void fadump_setup_crash_memory_ranges(void) static inline unsigned long fadump_relocate(unsigned long paddr) { if (paddr > RMA_START && paddr < fw_dump.boot_memory_size) - return fdm.rmr_region.destination_address + paddr; + return be64_to_cpu(fdm.rmr_region.destination_address) + paddr; else return paddr; } @@ -902,7 +902,7 @@ static int fadump_create_elfcore_headers(char *bufp) * to the specified destination_address. Hence set * the correct offset. */ - phdr->p_offset = fdm.rmr_region.destination_address; + phdr->p_offset = be64_to_cpu(fdm.rmr_region.destination_address); } phdr->p_paddr = mbase; @@ -951,7 +951,7 @@ static void register_fadump(void) fadump_setup_crash_memory_ranges(); - addr = fdm.rmr_region.destination_address + fdm.rmr_region.source_len; + addr = be64_to_cpu(fdm.rmr_region.destination_address) + be64_to_cpu(fdm.rmr_region.source_len); /* Initialize fadump crash info header. */ addr = init_fadump_header(addr); vaddr = __va(addr); @@ -1023,7 +1023,7 @@ void fadump_cleanup(void) /* Invalidate the registration only if dump is active. */ if (fw_dump.dump_active) { init_fadump_mem_struct(&fdm, - fdm_active->cpu_state_data.destination_address); + be64_to_cpu(fdm_active->cpu_state_data.destination_address)); fadump_invalidate_dump(&fdm); } } @@ -1063,7 +1063,7 @@ static void fadump_invalidate_release_mem(void) return; } - destination_address = fdm_active->cpu_state_data.destination_address; + destination_address = be64_to_cpu(fdm_active->cpu_state_data.destination_address); fadump_cleanup(); mutex_unlock(&fadump_mutex); @@ -1183,31 +1183,31 @@ static int fadump_region_show(struct seq_file *m, void *private) seq_printf(m, "CPU : [%#016llx-%#016llx] %#llx bytes, " "Dumped: %#llx\n", - fdm_ptr->cpu_state_data.destination_address, - fdm_ptr->cpu_state_data.destination_address + - fdm_ptr->cpu_state_data.source_len - 1, - fdm_ptr->cpu_state_data.source_len, - fdm_ptr->cpu_state_data.bytes_dumped); + be64_to_cpu(fdm_ptr->cpu_state_data.destination_address), + be64_to_cpu(fdm_ptr->cpu_state_data.destination_address) + + be64_to_cpu(fdm_ptr->cpu_state_data.source_len) - 1, + be64_to_cpu(fdm_ptr->cpu_state_data.source_len), + be64_to_cpu(fdm_ptr->cpu_state_data.bytes_dumped)); seq_printf(m, "HPTE: [%#016llx-%#016llx] %#llx bytes, " "Dumped: %#llx\n", - fdm_ptr->hpte_region.destination_address, - fdm_ptr->hpte_region.destination_address + - fdm_ptr->hpte_region.source_len - 1, - fdm_ptr->hpte_region.source_len, - fdm_ptr->hpte_region.bytes_dumped); + be64_to_cpu(fdm_ptr->hpte_region.destination_address), + be64_to_cpu(fdm_ptr->hpte_region.destination_address) + + be64_to_cpu(fdm_ptr->hpte_region.source_len) - 1, + be64_to_cpu(fdm_ptr->hpte_region.source_len), + be64_to_cpu(fdm_ptr->hpte_region.bytes_dumped)); seq_printf(m, "DUMP: [%#016llx-%#016llx] %#llx bytes, " "Dumped: %#llx\n", - fdm_ptr->rmr_region.destination_address, - fdm_ptr->rmr_region.destination_address + - fdm_ptr->rmr_region.source_len - 1, - fdm_ptr->rmr_region.source_len, - fdm_ptr->rmr_region.bytes_dumped); + be64_to_cpu(fdm_ptr->rmr_region.destination_address), + be64_to_cpu(fdm_ptr->rmr_region.destination_address) + + be64_to_cpu(fdm_ptr->rmr_region.source_len) - 1, + be64_to_cpu(fdm_ptr->rmr_region.source_len), + be64_to_cpu(fdm_ptr->rmr_region.bytes_dumped)); if (!fdm_active || (fw_dump.reserve_dump_area_start == - fdm_ptr->cpu_state_data.destination_address)) + be64_to_cpu(fdm_ptr->cpu_state_data.destination_address))) goto out; /* Dump is active. Show reserved memory region. */ @@ -1215,10 +1215,10 @@ static int fadump_region_show(struct seq_file *m, void *private) " : [%#016llx-%#016llx] %#llx bytes, " "Dumped: %#llx\n", (unsigned long long)fw_dump.reserve_dump_area_start, - fdm_ptr->cpu_state_data.destination_address - 1, - fdm_ptr->cpu_state_data.destination_address - + be64_to_cpu(fdm_ptr->cpu_state_data.destination_address) - 1, + be64_to_cpu(fdm_ptr->cpu_state_data.destination_address) - fw_dump.reserve_dump_area_start, - fdm_ptr->cpu_state_data.destination_address - + be64_to_cpu(fdm_ptr->cpu_state_data.destination_address) - fw_dump.reserve_dump_area_start); out: if (fdm_active) diff --git a/arch/powerpc/mm/init_32.c b/arch/powerpc/mm/init_32.c index cad68ff8eca5322990cfa56ee46003ae530e1d39..415a51b028b952c14b4f1ad3c88dd79b73657142 100644 --- a/arch/powerpc/mm/init_32.c +++ b/arch/powerpc/mm/init_32.c @@ -103,7 +103,7 @@ unsigned long __max_low_memory = MAX_LOW_MEM; /* * Check for command-line options that affect what MMU_init will do. */ -void MMU_setup(void) +void __init MMU_setup(void) { /* Check for nobats option (used in mapin_ram). */ if (strstr(boot_command_line, "nobats")) { diff --git a/arch/powerpc/platforms/powernv/opal-lpc.c b/arch/powerpc/platforms/powernv/opal-lpc.c index ad4b31df779a389ceb83b1c62d77a850e9c1b749..e4169d68cb3289a98230a7547b0ef24e3fb5c260 100644 --- a/arch/powerpc/platforms/powernv/opal-lpc.c +++ b/arch/powerpc/platforms/powernv/opal-lpc.c @@ -216,14 +216,54 @@ static ssize_t lpc_debug_read(struct file *filp, char __user *ubuf, &data, len); if (rc) return -ENXIO; + + /* + * Now there is some trickery with the data returned by OPAL + * as it's the desired data right justified in a 32-bit BE + * word. + * + * This is a very bad interface and I'm to blame for it :-( + * + * So we can't just apply a 32-bit swap to what comes from OPAL, + * because user space expects the *bytes* to be in their proper + * respective positions (ie, LPC position). + * + * So what we really want to do here is to shift data right + * appropriately on a LE kernel. + * + * IE. If the LPC transaction has bytes B0, B1, B2 and B3 in that + * order, we have in memory written to by OPAL at the "data" + * pointer: + * + * Bytes: OPAL "data" LE "data" + * 32-bit: B0 B1 B2 B3 B0B1B2B3 B3B2B1B0 + * 16-bit: B0 B1 0000B0B1 B1B00000 + * 8-bit: B0 000000B0 B0000000 + * + * So a BE kernel will have the leftmost of the above in the MSB + * and rightmost in the LSB and can just then "cast" the u32 "data" + * down to the appropriate quantity and write it. + * + * However, an LE kernel can't. It doesn't need to swap because a + * load from data followed by a store to user are going to preserve + * the byte ordering which is the wire byte order which is what the + * user wants, but in order to "crop" to the right size, we need to + * shift right first. + */ switch(len) { case 4: rc = __put_user((u32)data, (u32 __user *)ubuf); break; case 2: +#ifdef __LITTLE_ENDIAN__ + data >>= 16; +#endif rc = __put_user((u16)data, (u16 __user *)ubuf); break; default: +#ifdef __LITTLE_ENDIAN__ + data >>= 24; +#endif rc = __put_user((u8)data, (u8 __user *)ubuf); break; } @@ -263,12 +303,31 @@ static ssize_t lpc_debug_write(struct file *filp, const char __user *ubuf, else if (todo > 1 && (pos & 1) == 0) len = 2; } + + /* + * Similarly to the read case, we have some trickery here but + * it's different to handle. We need to pass the value to OPAL in + * a register whose layout depends on the access size. We want + * to reproduce the memory layout of the user, however we aren't + * doing a load from user and a store to another memory location + * which would achieve that. Here we pass the value to OPAL via + * a register which is expected to contain the "BE" interpretation + * of the byte sequence. IE: for a 32-bit access, byte 0 should be + * in the MSB. So here we *do* need to byteswap on LE. + * + * User bytes: LE "data" OPAL "data" + * 32-bit: B0 B1 B2 B3 B3B2B1B0 B0B1B2B3 + * 16-bit: B0 B1 0000B1B0 0000B0B1 + * 8-bit: B0 000000B0 000000B0 + */ switch(len) { case 4: rc = __get_user(data, (u32 __user *)ubuf); + data = cpu_to_be32(data); break; case 2: rc = __get_user(data, (u16 __user *)ubuf); + data = cpu_to_be16(data); break; default: rc = __get_user(data, (u8 __user *)ubuf); diff --git a/arch/powerpc/platforms/pseries/dlpar.c b/arch/powerpc/platforms/pseries/dlpar.c index 6ad83bd11fe21d0aaa22ad0d9cd6e0d399bb945d..c22bb1b4beb878c50bbe2f3c8d4dba7fd5d142f1 100644 --- a/arch/powerpc/platforms/pseries/dlpar.c +++ b/arch/powerpc/platforms/pseries/dlpar.c @@ -382,7 +382,7 @@ static int dlpar_online_cpu(struct device_node *dn) BUG_ON(get_cpu_current_state(cpu) != CPU_STATE_OFFLINE); cpu_maps_update_done(); - rc = cpu_up(cpu); + rc = device_online(get_cpu_device(cpu)); if (rc) goto out; cpu_maps_update_begin(); @@ -467,7 +467,7 @@ static int dlpar_offline_cpu(struct device_node *dn) if (get_cpu_current_state(cpu) == CPU_STATE_ONLINE) { set_preferred_offline_state(cpu, CPU_STATE_OFFLINE); cpu_maps_update_done(); - rc = cpu_down(cpu); + rc = device_offline(get_cpu_device(cpu)); if (rc) goto out; cpu_maps_update_begin(); diff --git a/arch/powerpc/platforms/pseries/lpar.c b/arch/powerpc/platforms/pseries/lpar.c index 8c509d5397c6e9a6f54c19ec569612e0005f96e2..f6880d2a40fbeffed49375c1cfcb174ce47ccb86 100644 --- a/arch/powerpc/platforms/pseries/lpar.c +++ b/arch/powerpc/platforms/pseries/lpar.c @@ -43,6 +43,7 @@ #include #include #include +#include #include "pseries.h" @@ -247,8 +248,17 @@ static void pSeries_lpar_hptab_clear(void) } #ifdef __LITTLE_ENDIAN__ - /* Reset exceptions to big endian */ - if (firmware_has_feature(FW_FEATURE_SET_MODE)) { + /* + * Reset exceptions to big endian. + * + * FIXME this is a hack for kexec, we need to reset the exception + * endian before starting the new kernel and this is a convenient place + * to do it. + * + * This is also called on boot when a fadump happens. In that case we + * must not change the exception endian mode. + */ + if (firmware_has_feature(FW_FEATURE_SET_MODE) && !is_fadump_active()) { long rc; rc = pseries_big_endian_exceptions();