diff options
Diffstat (limited to 'arch/parisc/mm/init.c')
| -rw-r--r-- | arch/parisc/mm/init.c | 727 |
1 files changed, 339 insertions, 388 deletions
diff --git a/arch/parisc/mm/init.c b/arch/parisc/mm/init.c index b0f96c0e6316..14270715d754 100644 --- a/arch/parisc/mm/init.c +++ b/arch/parisc/mm/init.c @@ -1,3 +1,4 @@ +// SPDX-License-Identifier: GPL-2.0 /* * linux/arch/parisc/mm/init.c * @@ -13,51 +14,46 @@ #include <linux/module.h> #include <linux/mm.h> -#include <linux/bootmem.h> +#include <linux/memblock.h> #include <linux/gfp.h> #include <linux/delay.h> #include <linux/init.h> -#include <linux/pci.h> /* for hppa_dma_ops and pcxl_dma_ops */ #include <linux/initrd.h> #include <linux/swap.h> #include <linux/unistd.h> #include <linux/nodemask.h> /* for node_online_map */ -#include <linux/pagemap.h> /* for release_pages and page_cache_release */ +#include <linux/pagemap.h> /* for release_pages */ +#include <linux/compat.h> +#include <linux/execmem.h> #include <asm/pgalloc.h> -#include <asm/pgtable.h> #include <asm/tlb.h> #include <asm/pdc_chassis.h> #include <asm/mmzone.h> #include <asm/sections.h> +#include <asm/msgbuf.h> +#include <asm/sparsemem.h> +#include <asm/asm-offsets.h> +#include <asm/shmbuf.h> extern int data_start; +extern void parisc_kernel_start(void); /* Kernel entry point in head.S */ -#if PT_NLEVELS == 3 -/* NOTE: This layout exactly conforms to the hybrid L2/L3 page table layout - * with the first pmd adjacent to the pgd and below it. gcc doesn't actually - * guarantee that global objects will be laid out in memory in the same order - * as the order of declaration, so put these in different sections and use - * the linker script to order them. */ -pmd_t pmd0[PTRS_PER_PMD] __attribute__ ((__section__ (".data..vm0.pmd"), aligned(PAGE_SIZE))); +#if CONFIG_PGTABLE_LEVELS == 3 +pmd_t pmd0[PTRS_PER_PMD] __section(".data..vm0.pmd") __attribute__ ((aligned(PAGE_SIZE))); #endif -pgd_t swapper_pg_dir[PTRS_PER_PGD] __attribute__ ((__section__ (".data..vm0.pgd"), aligned(PAGE_SIZE))); -pte_t pg0[PT_INITIAL * PTRS_PER_PTE] __attribute__ ((__section__ (".data..vm0.pte"), aligned(PAGE_SIZE))); - -#ifdef CONFIG_DISCONTIGMEM -struct node_map_data node_data[MAX_NUMNODES] __read_mostly; -signed char pfnnid_map[PFNNID_MAP_MAX] __read_mostly; -#endif +pgd_t swapper_pg_dir[PTRS_PER_PGD] __section(".data..vm0.pgd") __attribute__ ((aligned(PAGE_SIZE))); +pte_t pg0[PT_INITIAL * PTRS_PER_PTE] __section(".data..vm0.pte") __attribute__ ((aligned(PAGE_SIZE))); static struct resource data_resource = { .name = "Kernel data", - .flags = IORESOURCE_BUSY | IORESOURCE_MEM, + .flags = IORESOURCE_BUSY | IORESOURCE_SYSTEM_RAM, }; static struct resource code_resource = { .name = "Kernel code", - .flags = IORESOURCE_BUSY | IORESOURCE_MEM, + .flags = IORESOURCE_BUSY | IORESOURCE_SYSTEM_RAM, }; static struct resource pdcdata_resource = { @@ -67,17 +63,17 @@ static struct resource pdcdata_resource = { .flags = IORESOURCE_BUSY | IORESOURCE_MEM, }; -static struct resource sysram_resources[MAX_PHYSMEM_RANGES] __read_mostly; +static struct resource sysram_resources[MAX_PHYSMEM_RANGES] __ro_after_init; /* The following array is initialized from the firmware specific * information retrieved in kernel/inventory.c. */ -physmem_range_t pmem_ranges[MAX_PHYSMEM_RANGES] __read_mostly; -int npmem_ranges __read_mostly; +physmem_range_t pmem_ranges[MAX_PHYSMEM_RANGES] __initdata; +int npmem_ranges __initdata; #ifdef CONFIG_64BIT -#define MAX_MEM (~0UL) +#define MAX_MEM (1UL << MAX_PHYSMEM_BITS) #else /* !CONFIG_64BIT */ #define MAX_MEM (3584U*1024U*1024U) #endif /* !CONFIG_64BIT */ @@ -115,12 +111,8 @@ static void __init mem_limit_func(void) static void __init setup_bootmem(void) { - unsigned long bootmap_size; unsigned long mem_max; - unsigned long bootmap_pages; - unsigned long bootmap_start_pfn; - unsigned long bootmap_pfn; -#ifndef CONFIG_DISCONTIGMEM +#ifndef CONFIG_SPARSEMEM physmem_range_t pmem_holes[MAX_PHYSMEM_RANGES - 1]; int npmem_holes; #endif @@ -138,23 +130,16 @@ static void __init setup_bootmem(void) int j; for (j = i; j > 0; j--) { - unsigned long tmp; - if (pmem_ranges[j-1].start_pfn < pmem_ranges[j].start_pfn) { break; } - tmp = pmem_ranges[j-1].start_pfn; - pmem_ranges[j-1].start_pfn = pmem_ranges[j].start_pfn; - pmem_ranges[j].start_pfn = tmp; - tmp = pmem_ranges[j-1].pages; - pmem_ranges[j-1].pages = pmem_ranges[j].pages; - pmem_ranges[j].pages = tmp; + swap(pmem_ranges[j-1], pmem_ranges[j]); } } -#ifndef CONFIG_DISCONTIGMEM +#ifndef CONFIG_SPARSEMEM /* * Throw out ranges that are too far apart (controlled by * MAX_GAP). @@ -166,7 +151,7 @@ static void __init setup_bootmem(void) pmem_ranges[i-1].pages) > MAX_GAP) { npmem_ranges = i; printk("Large gap in memory detected (%ld pages). " - "Consider turning on CONFIG_DISCONTIGMEM\n", + "Consider turning on CONFIG_SPARSEMEM\n", pmem_ranges[i].start_pfn - (pmem_ranges[i-1].start_pfn + pmem_ranges[i-1].pages)); @@ -175,33 +160,29 @@ static void __init setup_bootmem(void) } #endif - if (npmem_ranges > 1) { + /* Print the memory ranges */ + pr_info("Memory Ranges:\n"); - /* Print the memory ranges */ - - printk(KERN_INFO "Memory Ranges:\n"); + for (i = 0; i < npmem_ranges; i++) { + struct resource *res = &sysram_resources[i]; + unsigned long start; + unsigned long size; - for (i = 0; i < npmem_ranges; i++) { - unsigned long start; - unsigned long size; + size = (pmem_ranges[i].pages << PAGE_SHIFT); + start = (pmem_ranges[i].start_pfn << PAGE_SHIFT); + pr_info("%2d) Start 0x%016lx End 0x%016lx Size %6ld MB\n", + i, start, start + (size - 1), size >> 20); - size = (pmem_ranges[i].pages << PAGE_SHIFT); - start = (pmem_ranges[i].start_pfn << PAGE_SHIFT); - printk(KERN_INFO "%2d) Start 0x%016lx End 0x%016lx Size %6ld MB\n", - i,start, start + (size - 1), size >> 20); - } - } - - sysram_resource_count = npmem_ranges; - for (i = 0; i < sysram_resource_count; i++) { - struct resource *res = &sysram_resources[i]; + /* request memory resource */ res->name = "System RAM"; - res->start = pmem_ranges[i].start_pfn << PAGE_SHIFT; - res->end = res->start + (pmem_ranges[i].pages << PAGE_SHIFT)-1; - res->flags = IORESOURCE_MEM | IORESOURCE_BUSY; + res->start = start; + res->end = start + size - 1; + res->flags = IORESOURCE_SYSTEM_RAM | IORESOURCE_BUSY; request_resource(&iomem_resource, res); } + sysram_resource_count = npmem_ranges; + /* * For 32 bit kernels we limit the amount of memory we can * support, in order to preserve enough kernel address space @@ -235,9 +216,8 @@ static void __init setup_bootmem(void) printk(KERN_INFO "Total Memory: %ld MB\n",mem_max >> 20); -#ifndef CONFIG_DISCONTIGMEM +#ifndef CONFIG_SPARSEMEM /* Merge the ranges, keeping track of the holes */ - { unsigned long end_pfn; unsigned long hole_pages; @@ -260,53 +240,37 @@ static void __init setup_bootmem(void) } #endif - bootmap_pages = 0; - for (i = 0; i < npmem_ranges; i++) - bootmap_pages += bootmem_bootmap_pages(pmem_ranges[i].pages); - - bootmap_start_pfn = PAGE_ALIGN(__pa((unsigned long) &_end)) >> PAGE_SHIFT; - -#ifdef CONFIG_DISCONTIGMEM - for (i = 0; i < MAX_PHYSMEM_RANGES; i++) { - memset(NODE_DATA(i), 0, sizeof(pg_data_t)); - NODE_DATA(i)->bdata = &bootmem_node_data[i]; - } - memset(pfnnid_map, 0xff, sizeof(pfnnid_map)); - - for (i = 0; i < npmem_ranges; i++) { - node_set_state(i, N_NORMAL_MEMORY); - node_set_online(i); - } -#endif - /* * Initialize and free the full range of memory in each range. - * Note that the only writing these routines do are to the bootmap, - * and we've made sure to locate the bootmap properly so that they - * won't be writing over anything important. */ - bootmap_pfn = bootmap_start_pfn; max_pfn = 0; for (i = 0; i < npmem_ranges; i++) { unsigned long start_pfn; unsigned long npages; + unsigned long start; + unsigned long size; start_pfn = pmem_ranges[i].start_pfn; npages = pmem_ranges[i].pages; - bootmap_size = init_bootmem_node(NODE_DATA(i), - bootmap_pfn, - start_pfn, - (start_pfn + npages) ); - free_bootmem_node(NODE_DATA(i), - (start_pfn << PAGE_SHIFT), - (npages << PAGE_SHIFT) ); - bootmap_pfn += (bootmap_size + PAGE_SIZE - 1) >> PAGE_SHIFT; + start = start_pfn << PAGE_SHIFT; + size = npages << PAGE_SHIFT; + + /* add system RAM memblock */ + memblock_add(start, size); + if ((start_pfn + npages) > max_pfn) max_pfn = start_pfn + npages; } + /* + * We can't use memblock top-down allocations because we only + * created the initial mapping up to KERNEL_INITIAL_SIZE in + * the assembly bootup code. + */ + memblock_set_bottom_up(true); + /* IOMMU is always used to access "high mem" on those boxes * that can support enough mem that a PCI device couldn't * directly DMA to any physical addresses. @@ -314,31 +278,22 @@ static void __init setup_bootmem(void) */ max_low_pfn = max_pfn; - /* bootmap sizing messed up? */ - BUG_ON((bootmap_pfn - bootmap_start_pfn) != bootmap_pages); - /* reserve PAGE0 pdc memory, kernel text/data/bss & bootmap */ #define PDC_CONSOLE_IO_IODC_SIZE 32768 - reserve_bootmem_node(NODE_DATA(0), 0UL, - (unsigned long)(PAGE0->mem_free + - PDC_CONSOLE_IO_IODC_SIZE), BOOTMEM_DEFAULT); - reserve_bootmem_node(NODE_DATA(0), __pa((unsigned long)_text), - (unsigned long)(_end - _text), BOOTMEM_DEFAULT); - reserve_bootmem_node(NODE_DATA(0), (bootmap_start_pfn << PAGE_SHIFT), - ((bootmap_pfn - bootmap_start_pfn) << PAGE_SHIFT), - BOOTMEM_DEFAULT); + memblock_reserve(0UL, (unsigned long)(PAGE0->mem_free + + PDC_CONSOLE_IO_IODC_SIZE)); + memblock_reserve(__pa(KERNEL_BINARY_TEXT_START), + (unsigned long)(_end - KERNEL_BINARY_TEXT_START)); -#ifndef CONFIG_DISCONTIGMEM +#ifndef CONFIG_SPARSEMEM /* reserve the holes */ for (i = 0; i < npmem_holes; i++) { - reserve_bootmem_node(NODE_DATA(0), - (pmem_holes[i].start_pfn << PAGE_SHIFT), - (pmem_holes[i].pages << PAGE_SHIFT), - BOOTMEM_DEFAULT); + memblock_reserve((pmem_holes[i].start_pfn << PAGE_SHIFT), + (pmem_holes[i].pages << PAGE_SHIFT)); } #endif @@ -356,8 +311,7 @@ static void __init setup_bootmem(void) initrd_below_start_ok = 1; printk(KERN_INFO "initrd: reserving %08lx-%08lx (mem_max %08lx)\n", __pa(initrd_start), __pa(initrd_start) + initrd_reserve, mem_max); - reserve_bootmem_node(NODE_DATA(0), __pa(initrd_start), - initrd_reserve, BOOTMEM_DEFAULT); + memblock_reserve(__pa(initrd_start), initrd_reserve); } } #endif @@ -376,13 +330,20 @@ static void __init setup_bootmem(void) request_resource(res, &data_resource); } request_resource(&sysram_resources[0], &pdcdata_resource); + + /* Initialize Page Deallocation Table (PDT) and check for bad memory. */ + pdc_pdt_init(); + + memblock_allow_resize(); + memblock_dump_all(); } -static void __init map_pages(unsigned long start_vaddr, - unsigned long start_paddr, unsigned long size, - pgprot_t pgprot, int force) +static bool kernel_set_to_readonly; + +static void __ref map_pages(unsigned long start_vaddr, + unsigned long start_paddr, unsigned long size, + pgprot_t pgprot, int force) { - pgd_t *pg_dir; pmd_t *pmd; pte_t *pg_table; unsigned long end_paddr; @@ -394,97 +355,71 @@ static void __init map_pages(unsigned long start_vaddr, unsigned long vaddr; unsigned long ro_start; unsigned long ro_end; - unsigned long fv_addr; - unsigned long gw_addr; - extern const unsigned long fault_vector_20; - extern void * const linux_gateway_page; + unsigned long kernel_start, kernel_end; ro_start = __pa((unsigned long)_text); ro_end = __pa((unsigned long)&data_start); - fv_addr = __pa((unsigned long)&fault_vector_20) & PAGE_MASK; - gw_addr = __pa((unsigned long)&linux_gateway_page) & PAGE_MASK; + kernel_start = __pa((unsigned long)&__init_begin); + kernel_end = __pa((unsigned long)&_end); end_paddr = start_paddr + size; - pg_dir = pgd_offset_k(start_vaddr); - -#if PTRS_PER_PMD == 1 - start_pmd = 0; -#else + /* for 2-level configuration PTRS_PER_PMD is 0 so start_pmd will be 0 */ start_pmd = ((start_vaddr >> PMD_SHIFT) & (PTRS_PER_PMD - 1)); -#endif start_pte = ((start_vaddr >> PAGE_SHIFT) & (PTRS_PER_PTE - 1)); address = start_paddr; vaddr = start_vaddr; while (address < end_paddr) { -#if PTRS_PER_PMD == 1 - pmd = (pmd_t *)__pa(pg_dir); -#else - pmd = (pmd_t *)pgd_address(*pg_dir); - - /* - * pmd is physical at this point - */ - - if (!pmd) { - pmd = (pmd_t *) alloc_bootmem_low_pages_node(NODE_DATA(0), PAGE_SIZE << PMD_ORDER); - pmd = (pmd_t *) __pa(pmd); + pgd_t *pgd = pgd_offset_k(vaddr); + p4d_t *p4d = p4d_offset(pgd, vaddr); + pud_t *pud = pud_offset(p4d, vaddr); + +#if CONFIG_PGTABLE_LEVELS == 3 + if (pud_none(*pud)) { + pmd = memblock_alloc_or_panic(PAGE_SIZE << PMD_TABLE_ORDER, + PAGE_SIZE << PMD_TABLE_ORDER); + pud_populate(NULL, pud, pmd); } - - pgd_populate(NULL, pg_dir, __va(pmd)); #endif - pg_dir++; - - /* now change pmd to kernel virtual addresses */ - pmd = (pmd_t *)__va(pmd) + start_pmd; + pmd = pmd_offset(pud, vaddr); for (tmp1 = start_pmd; tmp1 < PTRS_PER_PMD; tmp1++, pmd++) { - - /* - * pg_table is physical at this point - */ - - pg_table = (pte_t *)pmd_address(*pmd); - if (!pg_table) { - pg_table = (pte_t *) - alloc_bootmem_low_pages_node(NODE_DATA(0), PAGE_SIZE); - pg_table = (pte_t *) __pa(pg_table); + if (pmd_none(*pmd)) { + pg_table = memblock_alloc_or_panic(PAGE_SIZE, PAGE_SIZE); + pmd_populate_kernel(NULL, pmd, pg_table); } - pmd_populate_kernel(NULL, pmd, __va(pg_table)); - - /* now change pg_table to kernel virtual addresses */ - - pg_table = (pte_t *) __va(pg_table) + start_pte; + pg_table = pte_offset_kernel(pmd, vaddr); for (tmp2 = start_pte; tmp2 < PTRS_PER_PTE; tmp2++, pg_table++) { pte_t pte; + pgprot_t prot; + bool huge = false; + + if (force) { + prot = pgprot; + } else if (address < kernel_start || address >= kernel_end) { + /* outside kernel memory */ + prot = PAGE_KERNEL; + } else if (!kernel_set_to_readonly) { + /* still initializing, allow writing to RO memory */ + prot = PAGE_KERNEL_RWX; + huge = true; + } else if (address >= ro_start) { + /* Code (ro) and Data areas */ + prot = (address < ro_end) ? + PAGE_KERNEL_EXEC : PAGE_KERNEL; + huge = true; + } else { + prot = PAGE_KERNEL; + } - /* - * Map the fault vector writable so we can - * write the HPMC checksum. - */ - if (force) - pte = __mk_pte(address, pgprot); - else if (core_kernel_text(vaddr) && - address != fv_addr) - pte = __mk_pte(address, PAGE_KERNEL_EXEC); - else -#if defined(CONFIG_PARISC_PAGE_SIZE_4KB) - if (address >= ro_start && address < ro_end - && address != fv_addr - && address != gw_addr) - pte = __mk_pte(address, PAGE_KERNEL_RO); - else -#endif - pte = __mk_pte(address, pgprot); + pte = __mk_pte(address, prot); + if (huge) + pte = pte_mkhuge(pte); - if (address >= end_paddr) { - if (force) - break; - else - pte_val(pte) = 0; - } + if (address >= end_paddr) + break; set_pte(pg_table, pte); @@ -500,15 +435,32 @@ static void __init map_pages(unsigned long start_vaddr, } } +void __init set_kernel_text_rw(int enable_read_write) +{ + unsigned long start = (unsigned long) __init_begin; + unsigned long end = (unsigned long) &data_start; + + map_pages(start, __pa(start), end-start, + PAGE_KERNEL_RWX, enable_read_write ? 1:0); + + /* force the kernel to see the new page table entries */ + flush_cache_all(); + flush_tlb_all(); +} + void free_initmem(void) { unsigned long init_begin = (unsigned long)__init_begin; unsigned long init_end = (unsigned long)__init_end; + unsigned long kernel_end = (unsigned long)&_end; + + /* Remap kernel text and data, but do not touch init section yet. */ + map_pages(init_end, __pa(init_end), kernel_end - init_end, + PAGE_KERNEL, 0); /* The init text pages are marked R-X. We have to * flush the icache and mark them RW- * - * This is tricky, because map_pages is in the init section. * Do a dummy remap of the data section first (the data * section is already PAGE_KERNEL) to pull in the TLB entries * for map_kernel */ @@ -520,29 +472,34 @@ void free_initmem(void) PAGE_KERNEL, 1); /* force the kernel to see the new TLB entries */ - __flush_tlb_range(0, init_begin, init_end); - /* Attempt to catch anyone trying to execute code here - * by filling the page with BRK insns. - */ - memset((void *)init_begin, 0x00, init_end - init_begin); + __flush_tlb_range(0, init_begin, kernel_end); + /* finally dump all the instructions which were cached, since the * pages are no-longer executable */ flush_icache_range(init_begin, init_end); - - free_initmem_default(-1); + + free_initmem_default(POISON_FREE_INITMEM); /* set up a new led state on systems shipped LED State panel */ pdc_chassis_send_status(PDC_CHASSIS_DIRECT_BCOMPLETE); } -#ifdef CONFIG_DEBUG_RODATA +#ifdef CONFIG_STRICT_KERNEL_RWX void mark_rodata_ro(void) { - /* rodata memory was already mapped with KERNEL_RO access rights by - pagetable_init() and map_pages(). No need to do additional stuff here */ - printk (KERN_INFO "Write protecting the kernel read-only data: %luk\n", - (unsigned long)(__end_rodata - __start_rodata) >> 10); + unsigned long start = (unsigned long) &__start_rodata; + unsigned long end = (unsigned long) &__end_rodata; + + pr_info("Write protecting the kernel read-only data: %luk\n", + (end - start) >> 10); + + kernel_set_to_readonly = true; + map_pages(start, __pa(start), end - start, PAGE_KERNEL, 0); + + /* force the kernel to see the new page table entries */ + flush_cache_all(); + flush_tlb_all(); } #endif @@ -568,51 +525,71 @@ void mark_rodata_ro(void) #define SET_MAP_OFFSET(x) ((void *)(((unsigned long)(x) + VM_MAP_OFFSET) \ & ~(VM_MAP_OFFSET-1))) -void *parisc_vmalloc_start __read_mostly; +void *parisc_vmalloc_start __ro_after_init; EXPORT_SYMBOL(parisc_vmalloc_start); -#ifdef CONFIG_PA11 -unsigned long pcxl_dma_start __read_mostly; -#endif - void __init mem_init(void) { + /* Do sanity checks on IPC (compat) structures */ + BUILD_BUG_ON(sizeof(struct ipc64_perm) != 48); +#ifndef CONFIG_64BIT + BUILD_BUG_ON(sizeof(struct semid64_ds) != 80); + BUILD_BUG_ON(sizeof(struct msqid64_ds) != 104); + BUILD_BUG_ON(sizeof(struct shmid64_ds) != 104); +#endif +#ifdef CONFIG_COMPAT + BUILD_BUG_ON(sizeof(struct compat_ipc64_perm) != sizeof(struct ipc64_perm)); + BUILD_BUG_ON(sizeof(struct compat_semid64_ds) != 80); + BUILD_BUG_ON(sizeof(struct compat_msqid64_ds) != 104); + BUILD_BUG_ON(sizeof(struct compat_shmid64_ds) != 104); +#endif + /* Do sanity checks on page table constants */ BUILD_BUG_ON(PTE_ENTRY_SIZE != sizeof(pte_t)); BUILD_BUG_ON(PMD_ENTRY_SIZE != sizeof(pmd_t)); BUILD_BUG_ON(PGD_ENTRY_SIZE != sizeof(pgd_t)); BUILD_BUG_ON(PAGE_SHIFT + BITS_PER_PTE + BITS_PER_PMD + BITS_PER_PGD > BITS_PER_LONG); +#if CONFIG_PGTABLE_LEVELS == 3 + BUILD_BUG_ON(PT_INITIAL > PTRS_PER_PMD); +#else + BUILD_BUG_ON(PT_INITIAL > PTRS_PER_PGD); +#endif - high_memory = __va((max_pfn << PAGE_SHIFT)); - set_max_mapnr(page_to_pfn(virt_to_page(high_memory - 1)) + 1); - free_all_bootmem(); +#ifdef CONFIG_64BIT + /* avoid ldil_%L() asm statements to sign-extend into upper 32-bits */ + BUILD_BUG_ON(__PAGE_OFFSET >= 0x80000000); + BUILD_BUG_ON(TMPALIAS_MAP_START >= 0x80000000); +#endif #ifdef CONFIG_PA11 - if (hppa_dma_ops == &pcxl_dma_ops) { + if (boot_cpu_data.cpu_type == pcxl2 || boot_cpu_data.cpu_type == pcxl) { pcxl_dma_start = (unsigned long)SET_MAP_OFFSET(MAP_START); parisc_vmalloc_start = SET_MAP_OFFSET(pcxl_dma_start + PCXL_DMA_MAP_SIZE); - } else { - pcxl_dma_start = 0; - parisc_vmalloc_start = SET_MAP_OFFSET(MAP_START); - } -#else - parisc_vmalloc_start = SET_MAP_OFFSET(MAP_START); + } else #endif + parisc_vmalloc_start = SET_MAP_OFFSET(MAP_START); - mem_init_print_info(NULL); -#ifdef CONFIG_DEBUG_KERNEL /* double-sanity-check paranoia */ +#if 0 + /* + * Do not expose the virtual kernel memory layout to userspace. + * But keep code for debugging purposes. + */ printk("virtual kernel memory layout:\n" - " vmalloc : 0x%p - 0x%p (%4ld MB)\n" - " memory : 0x%p - 0x%p (%4ld MB)\n" - " .init : 0x%p - 0x%p (%4ld kB)\n" - " .data : 0x%p - 0x%p (%4ld kB)\n" - " .text : 0x%p - 0x%p (%4ld kB)\n", + " vmalloc : 0x%px - 0x%px (%4ld MB)\n" + " fixmap : 0x%px - 0x%px (%4ld kB)\n" + " memory : 0x%px - 0x%px (%4ld MB)\n" + " .init : 0x%px - 0x%px (%4ld kB)\n" + " .data : 0x%px - 0x%px (%4ld kB)\n" + " .text : 0x%px - 0x%px (%4ld kB)\n", (void*)VMALLOC_START, (void*)VMALLOC_END, (VMALLOC_END - VMALLOC_START) >> 20, + (void *)FIXMAP_START, (void *)(FIXMAP_START + FIXMAP_SIZE), + (unsigned long)(FIXMAP_SIZE / 1024), + __va(0), high_memory, ((unsigned long)high_memory - (unsigned long)__va(0)) >> 20, @@ -627,83 +604,9 @@ void __init mem_init(void) #endif } -unsigned long *empty_zero_page __read_mostly; +unsigned long *empty_zero_page __ro_after_init; EXPORT_SYMBOL(empty_zero_page); -void show_mem(unsigned int filter) -{ - int i,free = 0,total = 0,reserved = 0; - int shared = 0, cached = 0; - - printk(KERN_INFO "Mem-info:\n"); - show_free_areas(filter); - if (filter & SHOW_MEM_FILTER_PAGE_COUNT) - return; -#ifndef CONFIG_DISCONTIGMEM - i = max_mapnr; - while (i-- > 0) { - total++; - if (PageReserved(mem_map+i)) - reserved++; - else if (PageSwapCache(mem_map+i)) - cached++; - else if (!page_count(&mem_map[i])) - free++; - else - shared += page_count(&mem_map[i]) - 1; - } -#else - for (i = 0; i < npmem_ranges; i++) { - int j; - - for (j = node_start_pfn(i); j < node_end_pfn(i); j++) { - struct page *p; - unsigned long flags; - - pgdat_resize_lock(NODE_DATA(i), &flags); - p = nid_page_nr(i, j) - node_start_pfn(i); - - total++; - if (PageReserved(p)) - reserved++; - else if (PageSwapCache(p)) - cached++; - else if (!page_count(p)) - free++; - else - shared += page_count(p) - 1; - pgdat_resize_unlock(NODE_DATA(i), &flags); - } - } -#endif - printk(KERN_INFO "%d pages of RAM\n", total); - printk(KERN_INFO "%d reserved pages\n", reserved); - printk(KERN_INFO "%d pages shared\n", shared); - printk(KERN_INFO "%d pages swap cached\n", cached); - - -#ifdef CONFIG_DISCONTIGMEM - { - struct zonelist *zl; - int i, j; - - for (i = 0; i < npmem_ranges; i++) { - zl = node_zonelist(i, 0); - for (j = 0; j < MAX_NR_ZONES; j++) { - struct zoneref *z; - struct zone *zone; - - printk("Zone list for zone %d on node %d: ", j, i); - for_each_zone_zonelist(zone, z, zl, j) - printk("[%d/%s] ", zone_to_nid(zone), - zone->name); - printk("\n"); - } - } - } -#endif -} - /* * pagetable_init() sets up the page tables * @@ -720,11 +623,9 @@ static void __init pagetable_init(void) for (range = 0; range < npmem_ranges; range++) { unsigned long start_paddr; - unsigned long end_paddr; unsigned long size; start_paddr = pmem_ranges[range].start_pfn << PAGE_SHIFT; - end_paddr = start_paddr + (pmem_ranges[range].pages << PAGE_SHIFT); size = pmem_ranges[range].pages << PAGE_SHIFT; map_pages((unsigned long)__va(start_paddr), start_paddr, @@ -739,8 +640,8 @@ static void __init pagetable_init(void) } #endif - empty_zero_page = alloc_bootmem_pages(PAGE_SIZE); - memset(empty_zero_page, 0, PAGE_SIZE); + empty_zero_page = memblock_alloc_or_panic(PAGE_SIZE, PAGE_SIZE); + } static void __init gateway_init(void) @@ -763,109 +664,126 @@ static void __init gateway_init(void) PAGE_SIZE, PAGE_GATEWAY, 1); } -#ifdef CONFIG_HPUX -void -map_hpux_gateway_page(struct task_struct *tsk, struct mm_struct *mm) +static void __init fixmap_init(void) { - pgd_t *pg_dir; + unsigned long addr = FIXMAP_START; + unsigned long end = FIXMAP_START + FIXMAP_SIZE; + pgd_t *pgd = pgd_offset_k(addr); + p4d_t *p4d = p4d_offset(pgd, addr); + pud_t *pud = pud_offset(p4d, addr); pmd_t *pmd; - pte_t *pg_table; - unsigned long start_pmd; - unsigned long start_pte; - unsigned long address; - unsigned long hpux_gw_page_addr; - /* FIXME: This is 'const' in order to trick the compiler - into not treating it as DP-relative data. */ - extern void * const hpux_gateway_page; - hpux_gw_page_addr = HPUX_GATEWAY_ADDR & PAGE_MASK; + BUILD_BUG_ON(FIXMAP_SIZE > PMD_SIZE); - /* - * Setup HP-UX Gateway page. - * - * The HP-UX gateway page resides in the user address space, - * so it needs to be aliased into each process. - */ - - pg_dir = pgd_offset(mm,hpux_gw_page_addr); - -#if PTRS_PER_PMD == 1 - start_pmd = 0; -#else - start_pmd = ((hpux_gw_page_addr >> PMD_SHIFT) & (PTRS_PER_PMD - 1)); -#endif - start_pte = ((hpux_gw_page_addr >> PAGE_SHIFT) & (PTRS_PER_PTE - 1)); - - address = __pa(&hpux_gateway_page); -#if PTRS_PER_PMD == 1 - pmd = (pmd_t *)__pa(pg_dir); -#else - pmd = (pmd_t *) pgd_address(*pg_dir); - - /* - * pmd is physical at this point - */ - - if (!pmd) { - pmd = (pmd_t *) get_zeroed_page(GFP_KERNEL); - pmd = (pmd_t *) __pa(pmd); +#if CONFIG_PGTABLE_LEVELS == 3 + if (pud_none(*pud)) { + pmd = memblock_alloc_or_panic(PAGE_SIZE << PMD_TABLE_ORDER, + PAGE_SIZE << PMD_TABLE_ORDER); + pud_populate(NULL, pud, pmd); } - - __pgd_val_set(*pg_dir, PxD_FLAG_PRESENT | PxD_FLAG_VALID | (unsigned long) pmd); #endif - /* now change pmd to kernel virtual addresses */ - pmd = (pmd_t *)__va(pmd) + start_pmd; + pmd = pmd_offset(pud, addr); + do { + pte_t *pte = memblock_alloc_or_panic(PAGE_SIZE, PAGE_SIZE); - /* - * pg_table is physical at this point - */ + pmd_populate_kernel(&init_mm, pmd, pte); - pg_table = (pte_t *) pmd_address(*pmd); - if (!pg_table) - pg_table = (pte_t *) __pa(get_zeroed_page(GFP_KERNEL)); + addr += PAGE_SIZE; + } while (addr < end); +} - __pmd_val_set(*pmd, PxD_FLAG_PRESENT | PxD_FLAG_VALID | (unsigned long) pg_table); +static void __init parisc_bootmem_free(void) +{ + unsigned long max_zone_pfn[MAX_NR_ZONES] = { 0, }; - /* now change pg_table to kernel virtual addresses */ + max_zone_pfn[0] = memblock_end_of_DRAM(); - pg_table = (pte_t *) __va(pg_table) + start_pte; - set_pte(pg_table, __mk_pte(address, PAGE_GATEWAY)); + free_area_init(max_zone_pfn); } -EXPORT_SYMBOL(map_hpux_gateway_page); -#endif void __init paging_init(void) { - int i; - setup_bootmem(); pagetable_init(); gateway_init(); + fixmap_init(); flush_cache_all_local(); /* start with known state */ flush_tlb_all_local(NULL); - for (i = 0; i < npmem_ranges; i++) { - unsigned long zones_size[MAX_NR_ZONES] = { 0, }; - - zones_size[ZONE_NORMAL] = pmem_ranges[i].pages; - -#ifdef CONFIG_DISCONTIGMEM - /* Need to initialize the pfnnid_map before we can initialize - the zone */ - { - int j; - for (j = (pmem_ranges[i].start_pfn >> PFNNID_SHIFT); - j <= ((pmem_ranges[i].start_pfn + pmem_ranges[i].pages) >> PFNNID_SHIFT); - j++) { - pfnnid_map[j] = i; - } + sparse_init(); + parisc_bootmem_free(); +} + +static void alloc_btlb(unsigned long start, unsigned long end, int *slot, + unsigned long entry_info) +{ + const int slot_max = btlb_info.fixed_range_info.num_comb; + int min_num_pages = btlb_info.min_size; + unsigned long size; + + /* map at minimum 4 pages */ + if (min_num_pages < 4) + min_num_pages = 4; + + size = HUGEPAGE_SIZE; + while (start < end && *slot < slot_max && size >= PAGE_SIZE) { + /* starting address must have same alignment as size! */ + /* if correctly aligned and fits in double size, increase */ + if (((start & (2 * size - 1)) == 0) && + (end - start) >= (2 * size)) { + size <<= 1; + continue; } -#endif + /* if current size alignment is too big, try smaller size */ + if ((start & (size - 1)) != 0) { + size >>= 1; + continue; + } + if ((end - start) >= size) { + if ((size >> PAGE_SHIFT) >= min_num_pages) + pdc_btlb_insert(start >> PAGE_SHIFT, __pa(start) >> PAGE_SHIFT, + size >> PAGE_SHIFT, entry_info, *slot); + (*slot)++; + start += size; + continue; + } + size /= 2; + continue; + } +} - free_area_init_node(i, zones_size, - pmem_ranges[i].start_pfn, NULL); +void btlb_init_per_cpu(void) +{ + unsigned long s, t, e; + int slot; + + /* BTLBs are not available on 64-bit CPUs */ + if (IS_ENABLED(CONFIG_PA20)) + return; + else if (pdc_btlb_info(&btlb_info) < 0) { + memset(&btlb_info, 0, sizeof btlb_info); } + + /* insert BLTLBs for code and data segments */ + s = (uintptr_t) dereference_function_descriptor(&_stext); + e = (uintptr_t) dereference_function_descriptor(&_etext); + t = (uintptr_t) dereference_function_descriptor(&_sdata); + BUG_ON(t != e); + + /* code segments */ + slot = 0; + alloc_btlb(s, e, &slot, 0x13800000); + + /* sanity check */ + t = (uintptr_t) dereference_function_descriptor(&_edata); + e = (uintptr_t) dereference_function_descriptor(&__bss_start); + BUG_ON(t != e); + + /* data segments */ + s = (uintptr_t) dereference_function_descriptor(&_sdata); + e = (uintptr_t) dereference_function_descriptor(&__bss_stop); + alloc_btlb(s, e, &slot, 0x11800000); } #ifdef CONFIG_PA20 @@ -898,7 +816,7 @@ static unsigned long space_id[SID_ARRAY_SIZE] = { 1 }; /* disallow space 0 */ static unsigned long dirty_space_id[SID_ARRAY_SIZE]; static unsigned long space_id_index; static unsigned long free_space_ids = NR_SPACE_IDS - 1; -static unsigned long dirty_space_ids = 0; +static unsigned long dirty_space_ids; static DEFINE_SPINLOCK(sid_lock); @@ -920,7 +838,7 @@ unsigned long alloc_sid(void) free_space_ids--; index = find_next_zero_bit(space_id, NR_SPACE_IDS, space_id_index); - space_id[index >> SHIFT_PER_LONG] |= (1L << (index & (BITS_PER_LONG - 1))); + space_id[BIT_WORD(index)] |= BIT_MASK(index); space_id_index = index; spin_unlock(&sid_lock); @@ -931,16 +849,16 @@ unsigned long alloc_sid(void) void free_sid(unsigned long spaceid) { unsigned long index = spaceid >> SPACEID_SHIFT; - unsigned long *dirty_space_offset; + unsigned long *dirty_space_offset, mask; - dirty_space_offset = dirty_space_id + (index >> SHIFT_PER_LONG); - index &= (BITS_PER_LONG - 1); + dirty_space_offset = &dirty_space_id[BIT_WORD(index)]; + mask = BIT_MASK(index); spin_lock(&sid_lock); - BUG_ON(*dirty_space_offset & (1L << index)); /* attempt to free space id twice */ + BUG_ON(*dirty_space_offset & mask); /* attempt to free space id twice */ - *dirty_space_offset |= (1L << index); + *dirty_space_offset |= mask; dirty_space_ids++; spin_unlock(&sid_lock); @@ -1019,9 +937,9 @@ void flush_tlb_all(void) { int do_recycle; - __inc_irq_stat(irq_tlb_count); do_recycle = 0; spin_lock(&sid_lock); + __inc_irq_stat(irq_tlb_count); if (dirty_space_ids > RECYCLE_THRESHOLD) { BUG_ON(recycle_inuse); /* FIXME: Use a semaphore/wait queue here */ get_dirty_sids(&recycle_ndirty,recycle_dirty_array); @@ -1040,17 +958,50 @@ void flush_tlb_all(void) #else void flush_tlb_all(void) { - __inc_irq_stat(irq_tlb_count); spin_lock(&sid_lock); + __inc_irq_stat(irq_tlb_count); flush_tlb_all_local(NULL); recycle_sids(); spin_unlock(&sid_lock); } #endif -#ifdef CONFIG_BLK_DEV_INITRD -void free_initrd_mem(unsigned long start, unsigned long end) +static const pgprot_t protection_map[16] = { + [VM_NONE] = PAGE_NONE, + [VM_READ] = PAGE_READONLY, + [VM_WRITE] = PAGE_NONE, + [VM_WRITE | VM_READ] = PAGE_READONLY, + [VM_EXEC] = PAGE_EXECREAD, + [VM_EXEC | VM_READ] = PAGE_EXECREAD, + [VM_EXEC | VM_WRITE] = PAGE_EXECREAD, + [VM_EXEC | VM_WRITE | VM_READ] = PAGE_EXECREAD, + [VM_SHARED] = PAGE_NONE, + [VM_SHARED | VM_READ] = PAGE_READONLY, + [VM_SHARED | VM_WRITE] = PAGE_WRITEONLY, + [VM_SHARED | VM_WRITE | VM_READ] = PAGE_SHARED, + [VM_SHARED | VM_EXEC] = PAGE_EXECREAD, + [VM_SHARED | VM_EXEC | VM_READ] = PAGE_EXECREAD, + [VM_SHARED | VM_EXEC | VM_WRITE] = PAGE_RWX, + [VM_SHARED | VM_EXEC | VM_WRITE | VM_READ] = PAGE_RWX +}; +DECLARE_VM_GET_PAGE_PROT + +#ifdef CONFIG_EXECMEM +static struct execmem_info execmem_info __ro_after_init; + +struct execmem_info __init *execmem_arch_setup(void) { - free_reserved_area((void *)start, (void *)end, -1, "initrd"); + execmem_info = (struct execmem_info){ + .ranges = { + [EXECMEM_DEFAULT] = { + .start = VMALLOC_START, + .end = VMALLOC_END, + .pgprot = PAGE_KERNEL_RWX, + .alignment = 1, + }, + }, + }; + + return &execmem_info; } -#endif +#endif /* CONFIG_EXECMEM */ |