From 7c8c6b9776fb41134d87ef50706a777a45d61cd4 Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Thu, 6 Oct 2005 12:23:33 +1000 Subject: [PATCH] powerpc: Merge lmb.c and make MM initialization use it. This also creates merged versions of do_init_bootmem, paging_init and mem_init and moves them to arch/powerpc/mm/mem.c. It gets rid of the mem_pieces stuff. I made memory_limit a parameter to lmb_enforce_memory_limit rather than a global referenced by that function. This will require some small changes to ppc64 if we want to continue building ARCH=ppc64 using the merged lmb.c. Signed-off-by: Paul Mackerras --- arch/powerpc/mm/Makefile | 4 +- arch/powerpc/mm/init.c | 363 ++--------------------------------- arch/powerpc/mm/init64.c | 126 ------------ arch/powerpc/mm/lmb.c | 296 ++++++++++++++++++++++++++++ arch/powerpc/mm/mem.c | 185 +++++++++++++++++- arch/powerpc/mm/mem_pieces.c | 163 ---------------- arch/powerpc/mm/mem_pieces.h | 48 ----- arch/powerpc/mm/mmu_decl.h | 2 + arch/powerpc/mm/pgtable.c | 3 +- arch/powerpc/mm/ppc_mmu.c | 17 +- include/asm-ppc/page.h | 18 +- 11 files changed, 519 insertions(+), 706 deletions(-) create mode 100644 arch/powerpc/mm/lmb.c delete mode 100644 arch/powerpc/mm/mem_pieces.c delete mode 100644 arch/powerpc/mm/mem_pieces.h diff --git a/arch/powerpc/mm/Makefile b/arch/powerpc/mm/Makefile index 9f52c26acd86..afd3be112b79 100644 --- a/arch/powerpc/mm/Makefile +++ b/arch/powerpc/mm/Makefile @@ -2,9 +2,9 @@ # Makefile for the linux ppc-specific parts of the memory manager. # -obj-y := fault.o mem.o +obj-y := fault.o mem.o lmb.o obj-$(CONFIG_PPC32) += init.o pgtable.o mmu_context.o \ - mem_pieces.o tlb.o + tlb.o obj-$(CONFIG_PPC64) += init64.o pgtable64.o mmu_context64.o obj-$(CONFIG_PPC_STD_MMU_32) += ppc_mmu.o hash_32.o obj-$(CONFIG_40x) += 4xx_mmu.o diff --git a/arch/powerpc/mm/init.c b/arch/powerpc/mm/init.c index 3a81ef15c67e..bf13c14e66b3 100644 --- a/arch/powerpc/mm/init.c +++ b/arch/powerpc/mm/init.c @@ -45,8 +45,9 @@ #include #include #include +#include +#include -#include "mem_pieces.h" #include "mmu_decl.h" #if defined(CONFIG_KERNEL_START_BOOL) || defined(CONFIG_LOWMEM_SIZE_BOOL) @@ -65,17 +66,11 @@ unsigned long total_lowmem; unsigned long ppc_memstart; unsigned long ppc_memoffset = PAGE_OFFSET; -int mem_init_done; -int init_bootmem_done; int boot_mapsize; #ifdef CONFIG_PPC_PMAC unsigned long agp_special_page; #endif -extern char _end[]; -extern char etext[], _stext[]; -extern char __init_begin, __init_end; - #ifdef CONFIG_HIGHMEM pte_t *kmap_pte; pgprot_t kmap_prot; @@ -85,15 +80,15 @@ EXPORT_SYMBOL(kmap_pte); #endif void MMU_init(void); -void set_phys_avail(unsigned long total_ram); /* XXX should be in current.h -- paulus */ extern struct task_struct *current_set[NR_CPUS]; char *klimit = _end; -struct mem_pieces phys_avail; struct device_node *memory_node; +extern int init_bootmem_done; + /* * this tells the system to map all of ram with the segregs * (i.e. page tables) instead of the bats. @@ -102,84 +97,14 @@ struct device_node *memory_node; int __map_without_bats; int __map_without_ltlbs; -/* max amount of RAM to use */ -unsigned long __max_memory; /* max amount of low RAM to map in */ unsigned long __max_low_memory = MAX_LOW_MEM; /* - * Read in a property describing some pieces of memory. + * limit of what is accessible with initial MMU setup - + * 256MB usually, but only 16MB on 601. */ -static int __init get_mem_prop(char *name, struct mem_pieces *mp) -{ - struct reg_property *rp; - int i, s; - unsigned int *ip; - int nac = prom_n_addr_cells(memory_node); - int nsc = prom_n_size_cells(memory_node); - - ip = (unsigned int *) get_property(memory_node, name, &s); - if (ip == NULL) { - printk(KERN_ERR "error: couldn't get %s property on /memory\n", - name); - return 0; - } - s /= (nsc + nac) * 4; - rp = mp->regions; - for (i = 0; i < s; ++i, ip += nac+nsc) { - if (nac >= 2 && ip[nac-2] != 0) - continue; - rp->address = ip[nac-1]; - if (nsc >= 2 && ip[nac+nsc-2] != 0) - rp->size = ~0U; - else - rp->size = ip[nac+nsc-1]; - ++rp; - } - mp->n_regions = rp - mp->regions; - - /* Make sure the pieces are sorted. */ - mem_pieces_sort(mp); - mem_pieces_coalesce(mp); - return 1; -} - -/* - * Collect information about physical RAM and which pieces are - * already in use from the device tree. - */ -unsigned long __init find_end_of_memory(void) -{ - unsigned long a, total; - struct mem_pieces phys_mem; - - /* - * Find out where physical memory is, and check that it - * starts at 0 and is contiguous. It seems that RAM is - * always physically contiguous on Power Macintoshes. - * - * Supporting discontiguous physical memory isn't hard, - * it just makes the virtual <-> physical mapping functions - * more complicated (or else you end up wasting space - * in mem_map). - */ - memory_node = find_devices("memory"); - if (memory_node == NULL || !get_mem_prop("reg", &phys_mem) - || phys_mem.n_regions == 0) - panic("No RAM??"); - a = phys_mem.regions[0].address; - if (a != 0) - panic("RAM doesn't start at physical address 0"); - total = phys_mem.regions[0].size; - - if (phys_mem.n_regions > 1) { - printk("RAM starting at 0x%x is not contiguous\n", - phys_mem.regions[1].address); - printk("Using RAM from 0 to 0x%lx\n", total-1); - } - - return total; -} +unsigned long __initial_memory_limit = 0x10000000; /* * Check for command-line options that affect what MMU_init will do. @@ -194,27 +119,6 @@ void MMU_setup(void) if (strstr(cmd_line, "noltlbs")) { __map_without_ltlbs = 1; } - - /* Look for mem= option on command line */ - if (strstr(cmd_line, "mem=")) { - char *p, *q; - unsigned long maxmem = 0; - - for (q = cmd_line; (p = strstr(q, "mem=")) != 0; ) { - q = p + 4; - if (p > cmd_line && p[-1] != ' ') - continue; - maxmem = simple_strtoul(q, &q, 0); - if (*q == 'k' || *q == 'K') { - maxmem <<= 10; - ++q; - } else if (*q == 'm' || *q == 'M') { - maxmem <<= 20; - ++q; - } - } - __max_memory = maxmem; - } } /* @@ -227,23 +131,22 @@ void __init MMU_init(void) if (ppc_md.progress) ppc_md.progress("MMU:enter", 0x111); + /* 601 can only access 16MB at the moment */ + if (PVR_VER(mfspr(SPRN_PVR)) == 1) + __initial_memory_limit = 0x01000000; + /* parse args from command line */ MMU_setup(); - /* - * Figure out how much memory we have, how much - * is lowmem, and how much is highmem. If we were - * passed the total memory size from the bootloader, - * just use it. - */ - if (boot_mem_size) - total_memory = boot_mem_size; - else - total_memory = find_end_of_memory(); + if (lmb.memory.cnt > 1) { + lmb.memory.cnt = 1; + lmb_analyze(); + printk(KERN_WARNING "Only using first contiguous memory region"); + } - if (__max_memory && total_memory > __max_memory) - total_memory = __max_memory; + total_memory = lmb_end_of_DRAM(); total_lowmem = total_memory; + #ifdef CONFIG_FSL_BOOKE /* Freescale Book-E parts expect lowmem to be mapped by fixed TLB * entries, so we need to adjust lowmem to match the amount we can map @@ -256,7 +159,6 @@ void __init MMU_init(void) total_memory = total_lowmem; #endif /* CONFIG_HIGHMEM */ } - set_phys_avail(total_lowmem); /* Initialize the MMU hardware */ if (ppc_md.progress) @@ -303,7 +205,8 @@ void __init *early_get_page(void) if (init_bootmem_done) { p = alloc_bootmem_pages(PAGE_SIZE); } else { - p = mem_pieces_find(PAGE_SIZE, PAGE_SIZE); + p = __va(lmb_alloc_base(PAGE_SIZE, PAGE_SIZE, + __initial_memory_limit)); } return p; } @@ -353,229 +256,3 @@ void free_initrd_mem(unsigned long start, unsigned long end) } } #endif - -/* - * Initialize the bootmem system and give it all the memory we - * have available. - */ -void __init do_init_bootmem(void) -{ - unsigned long start, size; - int i; - - /* - * Find an area to use for the bootmem bitmap. - * We look for the first area which is at least - * 128kB in length (128kB is enough for a bitmap - * for 4GB of memory, using 4kB pages), plus 1 page - * (in case the address isn't page-aligned). - */ - start = 0; - size = 0; - for (i = 0; i < phys_avail.n_regions; ++i) { - unsigned long a = phys_avail.regions[i].address; - unsigned long s = phys_avail.regions[i].size; - if (s <= size) - continue; - start = a; - size = s; - if (s >= 33 * PAGE_SIZE) - break; - } - start = PAGE_ALIGN(start); - - min_low_pfn = start >> PAGE_SHIFT; - max_low_pfn = (PPC_MEMSTART + total_lowmem) >> PAGE_SHIFT; - max_pfn = (PPC_MEMSTART + total_memory) >> PAGE_SHIFT; - boot_mapsize = init_bootmem_node(&contig_page_data, min_low_pfn, - PPC_MEMSTART >> PAGE_SHIFT, - max_low_pfn); - - /* remove the bootmem bitmap from the available memory */ - mem_pieces_remove(&phys_avail, start, boot_mapsize, 1); - - /* add everything in phys_avail into the bootmem map */ - for (i = 0; i < phys_avail.n_regions; ++i) - free_bootmem(phys_avail.regions[i].address, - phys_avail.regions[i].size); - - init_bootmem_done = 1; -} - -/* - * paging_init() sets up the page tables - in fact we've already done this. - */ -void __init paging_init(void) -{ - unsigned long zones_size[MAX_NR_ZONES], i; - -#ifdef CONFIG_HIGHMEM - map_page(PKMAP_BASE, 0, 0); /* XXX gross */ - pkmap_page_table = pte_offset_kernel(pmd_offset(pgd_offset_k - (PKMAP_BASE), PKMAP_BASE), PKMAP_BASE); - map_page(KMAP_FIX_BEGIN, 0, 0); /* XXX gross */ - kmap_pte = pte_offset_kernel(pmd_offset(pgd_offset_k - (KMAP_FIX_BEGIN), KMAP_FIX_BEGIN), KMAP_FIX_BEGIN); - kmap_prot = PAGE_KERNEL; -#endif /* CONFIG_HIGHMEM */ - - /* - * All pages are DMA-able so we put them all in the DMA zone. - */ - zones_size[ZONE_DMA] = total_lowmem >> PAGE_SHIFT; - for (i = 1; i < MAX_NR_ZONES; i++) - zones_size[i] = 0; - -#ifdef CONFIG_HIGHMEM - zones_size[ZONE_HIGHMEM] = (total_memory - total_lowmem) >> PAGE_SHIFT; -#endif /* CONFIG_HIGHMEM */ - - free_area_init(zones_size); -} - -void __init mem_init(void) -{ - unsigned long addr; - int codepages = 0; - int datapages = 0; - int initpages = 0; -#ifdef CONFIG_HIGHMEM - unsigned long highmem_mapnr; - - highmem_mapnr = total_lowmem >> PAGE_SHIFT; -#endif /* CONFIG_HIGHMEM */ - max_mapnr = total_memory >> PAGE_SHIFT; - - high_memory = (void *) __va(PPC_MEMSTART + total_lowmem); - num_physpages = max_mapnr; /* RAM is assumed contiguous */ - - totalram_pages += free_all_bootmem(); - -#ifdef CONFIG_BLK_DEV_INITRD - /* if we are booted from BootX with an initial ramdisk, - make sure the ramdisk pages aren't reserved. */ - if (initrd_start) { - for (addr = initrd_start; addr < initrd_end; addr += PAGE_SIZE) - ClearPageReserved(virt_to_page(addr)); - } -#endif /* CONFIG_BLK_DEV_INITRD */ - -#ifdef CONFIG_PPC_OF - /* mark the RTAS pages as reserved */ - if ( rtas_data ) - for (addr = (ulong)__va(rtas_data); - addr < PAGE_ALIGN((ulong)__va(rtas_data)+rtas_size) ; - addr += PAGE_SIZE) - SetPageReserved(virt_to_page(addr)); -#endif -#ifdef CONFIG_PPC_PMAC - if (agp_special_page) - SetPageReserved(virt_to_page(agp_special_page)); -#endif - for (addr = PAGE_OFFSET; addr < (unsigned long)high_memory; - addr += PAGE_SIZE) { - if (!PageReserved(virt_to_page(addr))) - continue; - if (addr < (ulong) etext) - codepages++; - else if (addr >= (unsigned long)&__init_begin - && addr < (unsigned long)&__init_end) - initpages++; - else if (addr < (ulong) klimit) - datapages++; - } - -#ifdef CONFIG_HIGHMEM - { - unsigned long pfn; - - for (pfn = highmem_mapnr; pfn < max_mapnr; ++pfn) { - struct page *page = mem_map + pfn; - - ClearPageReserved(page); - set_page_count(page, 1); - __free_page(page); - totalhigh_pages++; - } - totalram_pages += totalhigh_pages; - } -#endif /* CONFIG_HIGHMEM */ - - printk("Memory: %luk available (%dk kernel code, %dk data, %dk init, %ldk highmem)\n", - (unsigned long)nr_free_pages()<< (PAGE_SHIFT-10), - codepages<< (PAGE_SHIFT-10), datapages<< (PAGE_SHIFT-10), - initpages<< (PAGE_SHIFT-10), - (unsigned long) (totalhigh_pages << (PAGE_SHIFT-10))); - -#ifdef CONFIG_PPC_PMAC - if (agp_special_page) - printk(KERN_INFO "AGP special page: 0x%08lx\n", agp_special_page); -#endif - - mem_init_done = 1; -} - -/* - * Set phys_avail to the amount of physical memory, - * less the kernel text/data/bss. - */ -void __init -set_phys_avail(unsigned long total_memory) -{ - unsigned long kstart, ksize; - - /* - * Initially, available physical memory is equivalent to all - * physical memory. - */ - - phys_avail.regions[0].address = PPC_MEMSTART; - phys_avail.regions[0].size = total_memory; - phys_avail.n_regions = 1; - - /* - * Map out the kernel text/data/bss from the available physical - * memory. - */ - - kstart = __pa(_stext); /* should be 0 */ - ksize = PAGE_ALIGN(klimit - _stext); - - mem_pieces_remove(&phys_avail, kstart, ksize, 0); - mem_pieces_remove(&phys_avail, 0, 0x4000, 0); - -#if defined(CONFIG_BLK_DEV_INITRD) - /* Remove the init RAM disk from the available memory. */ - if (initrd_start) { - mem_pieces_remove(&phys_avail, __pa(initrd_start), - initrd_end - initrd_start, 1); - } -#endif /* CONFIG_BLK_DEV_INITRD */ -#ifdef CONFIG_PPC_OF - /* remove the RTAS pages from the available memory */ - if (rtas_data) - mem_pieces_remove(&phys_avail, rtas_data, rtas_size, 1); -#endif -#ifdef CONFIG_PPC_PMAC - /* Because of some uninorth weirdness, we need a page of - * memory as high as possible (it must be outside of the - * bus address seen as the AGP aperture). It will be used - * by the r128 DRM driver - * - * FIXME: We need to make sure that page doesn't overlap any of the\ - * above. This could be done by improving mem_pieces_find to be able - * to do a backward search from the end of the list. - */ - if (_machine == _MACH_Pmac && find_devices("uni-north-agp")) { - agp_special_page = (total_memory - PAGE_SIZE); - mem_pieces_remove(&phys_avail, agp_special_page, PAGE_SIZE, 0); - agp_special_page = (unsigned long)__va(agp_special_page); - } -#endif /* CONFIG_PPC_PMAC */ -} - -/* Mark some memory as reserved by removing it from phys_avail. */ -void __init reserve_phys_mem(unsigned long start, unsigned long size) -{ - mem_pieces_remove(&phys_avail, start, size, 1); -} diff --git a/arch/powerpc/mm/init64.c b/arch/powerpc/mm/init64.c index 81f6745b31ef..c0ce6a7af3c7 100644 --- a/arch/powerpc/mm/init64.c +++ b/arch/powerpc/mm/init64.c @@ -166,77 +166,6 @@ void free_initrd_mem(unsigned long start, unsigned long end) } #endif -/* - * Initialize the bootmem system and give it all the memory we - * have available. - */ -#ifndef CONFIG_NEED_MULTIPLE_NODES -void __init do_init_bootmem(void) -{ - unsigned long i; - unsigned long start, bootmap_pages; - unsigned long total_pages = lmb_end_of_DRAM() >> PAGE_SHIFT; - int boot_mapsize; - - /* - * Find an area to use for the bootmem bitmap. Calculate the size of - * bitmap required as (Total Memory) / PAGE_SIZE / BITS_PER_BYTE. - * Add 1 additional page in case the address isn't page-aligned. - */ - bootmap_pages = bootmem_bootmap_pages(total_pages); - - start = lmb_alloc(bootmap_pages<> PAGE_SHIFT, total_pages); - - max_pfn = max_low_pfn; - - /* Add all physical memory to the bootmem map, mark each area - * present. - */ - for (i=0; i < lmb.memory.cnt; i++) - free_bootmem(lmb.memory.region[i].base, - lmb_size_bytes(&lmb.memory, i)); - - /* reserve the sections we're already using */ - for (i=0; i < lmb.reserved.cnt; i++) - reserve_bootmem(lmb.reserved.region[i].base, - lmb_size_bytes(&lmb.reserved, i)); - - for (i=0; i < lmb.memory.cnt; i++) - memory_present(0, lmb_start_pfn(&lmb.memory, i), - lmb_end_pfn(&lmb.memory, i)); -} - -/* - * paging_init() sets up the page tables - in fact we've already done this. - */ -void __init paging_init(void) -{ - unsigned long zones_size[MAX_NR_ZONES]; - unsigned long zholes_size[MAX_NR_ZONES]; - unsigned long total_ram = lmb_phys_mem_size(); - unsigned long top_of_ram = lmb_end_of_DRAM(); - - printk(KERN_INFO "Top of RAM: 0x%lx, Total RAM: 0x%lx\n", - top_of_ram, total_ram); - printk(KERN_INFO "Memory hole size: %ldMB\n", - (top_of_ram - total_ram) >> 20); - /* - * All pages are DMA-able so we put them all in the DMA zone. - */ - memset(zones_size, 0, sizeof(zones_size)); - memset(zholes_size, 0, sizeof(zholes_size)); - - zones_size[ZONE_DMA] = top_of_ram >> PAGE_SHIFT; - zholes_size[ZONE_DMA] = (top_of_ram - total_ram) >> PAGE_SHIFT; - - free_area_init_node(0, NODE_DATA(0), zones_size, - __pa(PAGE_OFFSET) >> PAGE_SHIFT, zholes_size); -} -#endif /* ! CONFIG_NEED_MULTIPLE_NODES */ - static struct kcore_list kcore_vmem; static int __init setup_kcore(void) @@ -264,61 +193,6 @@ static int __init setup_kcore(void) } module_init(setup_kcore); -void __init mem_init(void) -{ -#ifdef CONFIG_NEED_MULTIPLE_NODES - int nid; -#endif - pg_data_t *pgdat; - unsigned long i; - struct page *page; - unsigned long reservedpages = 0, codesize, initsize, datasize, bsssize; - - num_physpages = max_low_pfn; /* RAM is assumed contiguous */ - high_memory = (void *) __va(max_low_pfn * PAGE_SIZE); - -#ifdef CONFIG_NEED_MULTIPLE_NODES - for_each_online_node(nid) { - if (NODE_DATA(nid)->node_spanned_pages != 0) { - printk("freeing bootmem node %x\n", nid); - totalram_pages += - free_all_bootmem_node(NODE_DATA(nid)); - } - } -#else - max_mapnr = num_physpages; - totalram_pages += free_all_bootmem(); -#endif - - for_each_pgdat(pgdat) { - for (i = 0; i < pgdat->node_spanned_pages; i++) { - page = pgdat_page_nr(pgdat, i); - if (PageReserved(page)) - reservedpages++; - } - } - - codesize = (unsigned long)&_etext - (unsigned long)&_stext; - initsize = (unsigned long)&__init_end - (unsigned long)&__init_begin; - datasize = (unsigned long)&_edata - (unsigned long)&__init_end; - bsssize = (unsigned long)&__bss_stop - (unsigned long)&__bss_start; - - printk(KERN_INFO "Memory: %luk/%luk available (%luk kernel code, " - "%luk reserved, %luk data, %luk bss, %luk init)\n", - (unsigned long)nr_free_pages() << (PAGE_SHIFT-10), - num_physpages << (PAGE_SHIFT-10), - codesize >> 10, - reservedpages << (PAGE_SHIFT-10), - datasize >> 10, - bsssize >> 10, - initsize >> 10); - - mem_init_done = 1; - - /* Initialize the vDSO */ - vdso_init(); -} - void __iomem * reserve_phb_iospace(unsigned long size) { void __iomem *virt_addr; diff --git a/arch/powerpc/mm/lmb.c b/arch/powerpc/mm/lmb.c new file mode 100644 index 000000000000..9b5aa6808eb8 --- /dev/null +++ b/arch/powerpc/mm/lmb.c @@ -0,0 +1,296 @@ +/* + * Procedures for maintaining information about logical memory blocks. + * + * Peter Bergner, IBM Corp. June 2001. + * Copyright (C) 2001 Peter Bergner. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#ifdef CONFIG_PPC32 +#include "mmu_decl.h" /* for __max_low_memory */ +#endif + +struct lmb lmb; + +#undef DEBUG + +void lmb_dump_all(void) +{ +#ifdef DEBUG + unsigned long i; + + udbg_printf("lmb_dump_all:\n"); + udbg_printf(" memory.cnt = 0x%lx\n", + lmb.memory.cnt); + udbg_printf(" memory.size = 0x%lx\n", + lmb.memory.size); + for (i=0; i < lmb.memory.cnt ;i++) { + udbg_printf(" memory.region[0x%x].base = 0x%lx\n", + i, lmb.memory.region[i].base); + udbg_printf(" .size = 0x%lx\n", + lmb.memory.region[i].size); + } + + udbg_printf("\n reserved.cnt = 0x%lx\n", + lmb.reserved.cnt); + udbg_printf(" reserved.size = 0x%lx\n", + lmb.reserved.size); + for (i=0; i < lmb.reserved.cnt ;i++) { + udbg_printf(" reserved.region[0x%x].base = 0x%lx\n", + i, lmb.reserved.region[i].base); + udbg_printf(" .size = 0x%lx\n", + lmb.reserved.region[i].size); + } +#endif /* DEBUG */ +} + +static unsigned long __init lmb_addrs_overlap(unsigned long base1, + unsigned long size1, unsigned long base2, unsigned long size2) +{ + return ((base1 < (base2+size2)) && (base2 < (base1+size1))); +} + +static long __init lmb_addrs_adjacent(unsigned long base1, unsigned long size1, + unsigned long base2, unsigned long size2) +{ + if (base2 == base1 + size1) + return 1; + else if (base1 == base2 + size2) + return -1; + + return 0; +} + +static long __init lmb_regions_adjacent(struct lmb_region *rgn, + unsigned long r1, unsigned long r2) +{ + unsigned long base1 = rgn->region[r1].base; + unsigned long size1 = rgn->region[r1].size; + unsigned long base2 = rgn->region[r2].base; + unsigned long size2 = rgn->region[r2].size; + + return lmb_addrs_adjacent(base1, size1, base2, size2); +} + +/* Assumption: base addr of region 1 < base addr of region 2 */ +static void __init lmb_coalesce_regions(struct lmb_region *rgn, + unsigned long r1, unsigned long r2) +{ + unsigned long i; + + rgn->region[r1].size += rgn->region[r2].size; + for (i=r2; i < rgn->cnt-1; i++) { + rgn->region[i].base = rgn->region[i+1].base; + rgn->region[i].size = rgn->region[i+1].size; + } + rgn->cnt--; +} + +/* This routine called with relocation disabled. */ +void __init lmb_init(void) +{ + /* Create a dummy zero size LMB which will get coalesced away later. + * This simplifies the lmb_add() code below... + */ + lmb.memory.region[0].base = 0; + lmb.memory.region[0].size = 0; + lmb.memory.cnt = 1; + + /* Ditto. */ + lmb.reserved.region[0].base = 0; + lmb.reserved.region[0].size = 0; + lmb.reserved.cnt = 1; +} + +/* This routine may be called with relocation disabled. */ +void __init lmb_analyze(void) +{ + int i; + + lmb.memory.size = 0; + + for (i = 0; i < lmb.memory.cnt; i++) + lmb.memory.size += lmb.memory.region[i].size; +} + +/* This routine called with relocation disabled. */ +static long __init lmb_add_region(struct lmb_region *rgn, unsigned long base, + unsigned long size) +{ + unsigned long i, coalesced = 0; + long adjacent; + + /* First try and coalesce this LMB with another. */ + for (i=0; i < rgn->cnt; i++) { + unsigned long rgnbase = rgn->region[i].base; + unsigned long rgnsize = rgn->region[i].size; + + adjacent = lmb_addrs_adjacent(base,size,rgnbase,rgnsize); + if ( adjacent > 0 ) { + rgn->region[i].base -= size; + rgn->region[i].size += size; + coalesced++; + break; + } + else if ( adjacent < 0 ) { + rgn->region[i].size += size; + coalesced++; + break; + } + } + + if ((i < rgn->cnt-1) && lmb_regions_adjacent(rgn, i, i+1) ) { + lmb_coalesce_regions(rgn, i, i+1); + coalesced++; + } + + if (coalesced) + return coalesced; + if (rgn->cnt >= MAX_LMB_REGIONS) + return -1; + + /* Couldn't coalesce the LMB, so add it to the sorted table. */ + for (i = rgn->cnt-1; i >= 0; i--) { + if (base < rgn->region[i].base) { + rgn->region[i+1].base = rgn->region[i].base; + rgn->region[i+1].size = rgn->region[i].size; + } else { + rgn->region[i+1].base = base; + rgn->region[i+1].size = size; + break; + } + } + rgn->cnt++; + + return 0; +} + +/* This routine may be called with relocation disabled. */ +long __init lmb_add(unsigned long base, unsigned long size) +{ + struct lmb_region *_rgn = &(lmb.memory); + + /* On pSeries LPAR systems, the first LMB is our RMO region. */ + if (base == 0) + lmb.rmo_size = size; + + return lmb_add_region(_rgn, base, size); + +} + +long __init lmb_reserve(unsigned long base, unsigned long size) +{ + struct lmb_region *_rgn = &(lmb.reserved); + + return lmb_add_region(_rgn, base, size); +} + +long __init lmb_overlaps_region(struct lmb_region *rgn, unsigned long base, + unsigned long size) +{ + unsigned long i; + + for (i=0; i < rgn->cnt; i++) { + unsigned long rgnbase = rgn->region[i].base; + unsigned long rgnsize = rgn->region[i].size; + if ( lmb_addrs_overlap(base,size,rgnbase,rgnsize) ) { + break; + } + } + + return (i < rgn->cnt) ? i : -1; +} + +unsigned long __init lmb_alloc(unsigned long size, unsigned long align) +{ + return lmb_alloc_base(size, align, LMB_ALLOC_ANYWHERE); +} + +unsigned long __init lmb_alloc_base(unsigned long size, unsigned long align, + unsigned long max_addr) +{ + long i, j; + unsigned long base = 0; + +#ifdef CONFIG_PPC32 + /* On 32-bit, make sure we allocate lowmem */ + if (max_addr == LMB_ALLOC_ANYWHERE) + max_addr = __max_low_memory; +#endif + for (i = lmb.memory.cnt-1; i >= 0; i--) { + unsigned long lmbbase = lmb.memory.region[i].base; + unsigned long lmbsize = lmb.memory.region[i].size; + + if (max_addr == LMB_ALLOC_ANYWHERE) + base = _ALIGN_DOWN(lmbbase + lmbsize - size, align); + else if (lmbbase < max_addr) { + base = min(lmbbase + lmbsize, max_addr); + base = _ALIGN_DOWN(base - size, align); + } else + continue; + + while ((lmbbase <= base) && + ((j = lmb_overlaps_region(&lmb.reserved, base, size)) >= 0) ) + base = _ALIGN_DOWN(lmb.reserved.region[j].base - size, + align); + + if ((base != 0) && (lmbbase <= base)) + break; + } + + if (i < 0) + return 0; + + lmb_add_region(&lmb.reserved, base, size); + + return base; +} + +/* You must call lmb_analyze() before this. */ +unsigned long __init lmb_phys_mem_size(void) +{ + return lmb.memory.size; +} + +unsigned long __init lmb_end_of_DRAM(void) +{ + int idx = lmb.memory.cnt - 1; + + return (lmb.memory.region[idx].base + lmb.memory.region[idx].size); +} + +/* + * Truncate the lmb list to memory_limit if it's set + * You must call lmb_analyze() after this. + */ +void __init lmb_enforce_memory_limit(unsigned long memory_limit) +{ + unsigned long i, limit; + + if (! memory_limit) + return; + + limit = memory_limit; + for (i = 0; i < lmb.memory.cnt; i++) { + if (limit > lmb.memory.region[i].size) { + limit -= lmb.memory.region[i].size; + continue; + } + + lmb.memory.region[i].size = limit; + lmb.memory.cnt = i + 1; + break; + } +} diff --git a/arch/powerpc/mm/mem.c b/arch/powerpc/mm/mem.c index 345db08e5d20..0650de74d0b3 100644 --- a/arch/powerpc/mm/mem.c +++ b/arch/powerpc/mm/mem.c @@ -45,8 +45,9 @@ #include #include #include +#include +#include -#include "mem_pieces.h" #include "mmu_decl.h" #ifndef CPU_FTR_COHERENT_ICACHE @@ -54,6 +55,9 @@ #define CPU_FTR_NOEXECUTE 0 #endif +int init_bootmem_done; +int mem_init_done; + /* * This is called by /dev/mem to know if a given address has to * be mapped non-cacheable or not @@ -130,6 +134,185 @@ void show_mem(void) printk("%ld pages swap cached\n", cached); } +/* + * Initialize the bootmem system and give it all the memory we + * have available. If we are using highmem, we only put the + * lowmem into the bootmem system. + */ +#ifndef CONFIG_NEED_MULTIPLE_NODES +void __init do_init_bootmem(void) +{ + unsigned long i; + unsigned long start, bootmap_pages; + unsigned long total_pages; + int boot_mapsize; + + max_pfn = total_pages = lmb_end_of_DRAM() >> PAGE_SHIFT; +#ifdef CONFIG_HIGHMEM + total_pages = total_lowmem >> PAGE_SHIFT; +#endif + + /* + * Find an area to use for the bootmem bitmap. Calculate the size of + * bitmap required as (Total Memory) / PAGE_SIZE / BITS_PER_BYTE. + * Add 1 additional page in case the address isn't page-aligned. + */ + bootmap_pages = bootmem_bootmap_pages(total_pages); + + start = lmb_alloc(bootmap_pages << PAGE_SHIFT, PAGE_SIZE); + BUG_ON(!start); + + boot_mapsize = init_bootmem(start >> PAGE_SHIFT, total_pages); + + /* Add all physical memory to the bootmem map, mark each area + * present. + */ + for (i = 0; i < lmb.memory.cnt; i++) { + unsigned long base = lmb.memory.region[i].base; + unsigned long size = lmb_size_bytes(&lmb.memory, i); +#ifdef CONFIG_HIGHMEM + if (base >= total_lowmem) + continue; + if (base + size > total_lowmem) + size = total_lowmem - base; +#endif + free_bootmem(base, size); + } + + /* reserve the sections we're already using */ + for (i = 0; i < lmb.reserved.cnt; i++) + reserve_bootmem(lmb.reserved.region[i].base, + lmb_size_bytes(&lmb.reserved, i)); + + /* XXX need to clip this if using highmem? */ + for (i = 0; i < lmb.memory.cnt; i++) + memory_present(0, lmb_start_pfn(&lmb.memory, i), + lmb_end_pfn(&lmb.memory, i)); + init_bootmem_done = 1; +} + +/* + * paging_init() sets up the page tables - in fact we've already done this. + */ +void __init paging_init(void) +{ + unsigned long zones_size[MAX_NR_ZONES]; + unsigned long zholes_size[MAX_NR_ZONES]; + unsigned long total_ram = lmb_phys_mem_size(); + unsigned long top_of_ram = lmb_end_of_DRAM(); + +#ifdef CONFIG_HIGHMEM + map_page(PKMAP_BASE, 0, 0); /* XXX gross */ + pkmap_page_table = pte_offset_kernel(pmd_offset(pgd_offset_k + (PKMAP_BASE), PKMAP_BASE), PKMAP_BASE); + map_page(KMAP_FIX_BEGIN, 0, 0); /* XXX gross */ + kmap_pte = pte_offset_kernel(pmd_offset(pgd_offset_k + (KMAP_FIX_BEGIN), KMAP_FIX_BEGIN), KMAP_FIX_BEGIN); + kmap_prot = PAGE_KERNEL; +#endif /* CONFIG_HIGHMEM */ + + printk(KERN_INFO "Top of RAM: 0x%lx, Total RAM: 0x%lx\n", + top_of_ram, total_ram); + printk(KERN_INFO "Memory hole size: %ldMB\n", + (top_of_ram - total_ram) >> 20); + /* + * All pages are DMA-able so we put them all in the DMA zone. + */ + memset(zones_size, 0, sizeof(zones_size)); + memset(zholes_size, 0, sizeof(zholes_size)); + + zones_size[ZONE_DMA] = top_of_ram >> PAGE_SHIFT; + zholes_size[ZONE_DMA] = (top_of_ram - total_ram) >> PAGE_SHIFT; + +#ifdef CONFIG_HIGHMEM + zones_size[ZONE_DMA] = total_lowmem >> PAGE_SHIFT; + zones_size[ZONE_HIGHMEM] = (total_memory - total_lowmem) >> PAGE_SHIFT; + zholes_size[ZONE_HIGHMEM] = (top_of_ram - total_ram) >> PAGE_SHIFT; +#else + zones_size[ZONE_DMA] = top_of_ram >> PAGE_SHIFT; + zholes_size[ZONE_DMA] = (top_of_ram - total_ram) >> PAGE_SHIFT; +#endif /* CONFIG_HIGHMEM */ + + free_area_init_node(0, NODE_DATA(0), zones_size, + __pa(PAGE_OFFSET) >> PAGE_SHIFT, zholes_size); +} +#endif /* ! CONFIG_NEED_MULTIPLE_NODES */ + +void __init mem_init(void) +{ +#ifdef CONFIG_NEED_MULTIPLE_NODES + int nid; +#endif + pg_data_t *pgdat; + unsigned long i; + struct page *page; + unsigned long reservedpages = 0, codesize, initsize, datasize, bsssize; + + num_physpages = max_pfn; /* RAM is assumed contiguous */ + high_memory = (void *) __va(max_low_pfn * PAGE_SIZE); + +#ifdef CONFIG_NEED_MULTIPLE_NODES + for_each_online_node(nid) { + if (NODE_DATA(nid)->node_spanned_pages != 0) { + printk("freeing bootmem node %x\n", nid); + totalram_pages += + free_all_bootmem_node(NODE_DATA(nid)); + } + } +#else + max_mapnr = num_physpages; + totalram_pages += free_all_bootmem(); +#endif + for_each_pgdat(pgdat) { + for (i = 0; i < pgdat->node_spanned_pages; i++) { + page = pgdat_page_nr(pgdat, i); + if (PageReserved(page)) + reservedpages++; + } + } + + codesize = (unsigned long)&_sdata - (unsigned long)&_stext; + datasize = (unsigned long)&__init_begin - (unsigned long)&_sdata; + initsize = (unsigned long)&__init_end - (unsigned long)&__init_begin; + bsssize = (unsigned long)&__bss_stop - (unsigned long)&__bss_start; + +#ifdef CONFIG_HIGHMEM + { + unsigned long pfn, highmem_mapnr; + + highmem_mapnr = total_lowmem >> PAGE_SHIFT; + for (pfn = highmem_mapnr; pfn < max_mapnr; ++pfn) { + struct page *page = pfn_to_page(pfn); + + ClearPageReserved(page); + set_page_count(page, 1); + __free_page(page); + totalhigh_pages++; + } + totalram_pages += totalhigh_pages; + printk(KERN_INFO "High memory: %luk\n", + totalhigh_pages << (PAGE_SHIFT-10)); + } +#endif /* CONFIG_HIGHMEM */ + + printk(KERN_INFO "Memory: %luk/%luk available (%luk kernel code, " + "%luk reserved, %luk data, %luk bss, %luk init)\n", + (unsigned long)nr_free_pages() << (PAGE_SHIFT-10), + num_physpages << (PAGE_SHIFT-10), + codesize >> 10, + reservedpages << (PAGE_SHIFT-10), + datasize >> 10, + bsssize >> 10, + initsize >> 10); + + mem_init_done = 1; + +#ifdef CONFIG_PPC64 + /* Initialize the vDSO */ + vdso_init(); +#endif +} + /* * This is called when a page has been modified by the kernel. * It just marks the page as not i-cache clean. We do the i-cache diff --git a/arch/powerpc/mm/mem_pieces.c b/arch/powerpc/mm/mem_pieces.c deleted file mode 100644 index 3d639052017e..000000000000 --- a/arch/powerpc/mm/mem_pieces.c +++ /dev/null @@ -1,163 +0,0 @@ -/* - * Copyright (c) 1996 Paul Mackerras - * Changes to accommodate Power Macintoshes. - * Cort Dougan - * Rewrites. - * Grant Erickson - * General rework and split from mm/init.c. - * - * Module name: mem_pieces.c - * - * Description: - * Routines and data structures for manipulating and representing - * phyiscal memory extents (i.e. address/length pairs). - * - */ - -#include -#include -#include -#include -#include - -#include "mem_pieces.h" - -extern struct mem_pieces phys_avail; - -static void mem_pieces_print(struct mem_pieces *); - -/* - * Scan a region for a piece of a given size with the required alignment. - */ -void __init * -mem_pieces_find(unsigned int size, unsigned int align) -{ - int i; - unsigned a, e; - struct mem_pieces *mp = &phys_avail; - - for (i = 0; i < mp->n_regions; ++i) { - a = mp->regions[i].address; - e = a + mp->regions[i].size; - a = (a + align - 1) & -align; - if (a + size <= e) { - mem_pieces_remove(mp, a, size, 1); - return (void *) __va(a); - } - } - panic("Couldn't find %u bytes at %u alignment\n", size, align); - - return NULL; -} - -/* - * Remove some memory from an array of pieces - */ -void __init -mem_pieces_remove(struct mem_pieces *mp, unsigned int start, unsigned int size, - int must_exist) -{ - int i, j; - unsigned int end, rs, re; - struct reg_property *rp; - - end = start + size; - for (i = 0, rp = mp->regions; i < mp->n_regions; ++i, ++rp) { - if (end > rp->address && start < rp->address + rp->size) - break; - } - if (i >= mp->n_regions) { - if (must_exist) - printk("mem_pieces_remove: [%x,%x) not in any region\n", - start, end); - return; - } - for (; i < mp->n_regions && end > rp->address; ++i, ++rp) { - rs = rp->address; - re = rs + rp->size; - if (must_exist && (start < rs || end > re)) { - printk("mem_pieces_remove: bad overlap [%x,%x) with", - start, end); - mem_pieces_print(mp); - must_exist = 0; - } - if (start > rs) { - rp->size = start - rs; - if (end < re) { - /* need to split this entry */ - if (mp->n_regions >= MEM_PIECES_MAX) - panic("eek... mem_pieces overflow"); - for (j = mp->n_regions; j > i + 1; --j) - mp->regions[j] = mp->regions[j-1]; - ++mp->n_regions; - rp[1].address = end; - rp[1].size = re - end; - } - } else { - if (end < re) { - rp->address = end; - rp->size = re - end; - } else { - /* need to delete this entry */ - for (j = i; j < mp->n_regions - 1; ++j) - mp->regions[j] = mp->regions[j+1]; - --mp->n_regions; - --i; - --rp; - } - } - } -} - -static void __init -mem_pieces_print(struct mem_pieces *mp) -{ - int i; - - for (i = 0; i < mp->n_regions; ++i) - printk(" [%x, %x)", mp->regions[i].address, - mp->regions[i].address + mp->regions[i].size); - printk("\n"); -} - -void __init -mem_pieces_sort(struct mem_pieces *mp) -{ - unsigned long a, s; - int i, j; - - for (i = 1; i < mp->n_regions; ++i) { - a = mp->regions[i].address; - s = mp->regions[i].size; - for (j = i - 1; j >= 0; --j) { - if (a >= mp->regions[j].address) - break; - mp->regions[j+1] = mp->regions[j]; - } - mp->regions[j+1].address = a; - mp->regions[j+1].size = s; - } -} - -void __init -mem_pieces_coalesce(struct mem_pieces *mp) -{ - unsigned long a, s, ns; - int i, j, d; - - d = 0; - for (i = 0; i < mp->n_regions; i = j) { - a = mp->regions[i].address; - s = mp->regions[i].size; - for (j = i + 1; j < mp->n_regions - && mp->regions[j].address - a <= s; ++j) { - ns = mp->regions[j].address + mp->regions[j].size - a; - if (ns > s) - s = ns; - } - mp->regions[d].address = a; - mp->regions[d].size = s; - ++d; - } - mp->n_regions = d; -} diff --git a/arch/powerpc/mm/mem_pieces.h b/arch/powerpc/mm/mem_pieces.h deleted file mode 100644 index e2b700dc7f18..000000000000 --- a/arch/powerpc/mm/mem_pieces.h +++ /dev/null @@ -1,48 +0,0 @@ -/* - * Copyright (c) 1996 Paul Mackerras - * Changes to accommodate Power Macintoshes. - * Cort Dougan - * Rewrites. - * Grant Erickson - * General rework and split from mm/init.c. - * - * Module name: mem_pieces.h - * - * Description: - * Routines and data structures for manipulating and representing - * phyiscal memory extents (i.e. address/length pairs). - * - */ - -#ifndef __MEM_PIECES_H__ -#define __MEM_PIECES_H__ - -#include - -#ifdef __cplusplus -extern "C" { -#endif - - -/* Type Definitions */ - -#define MEM_PIECES_MAX 32 - -struct mem_pieces { - int n_regions; - struct reg_property regions[MEM_PIECES_MAX]; -}; - -/* Function Prototypes */ - -extern void *mem_pieces_find(unsigned int size, unsigned int align); -extern void mem_pieces_remove(struct mem_pieces *mp, unsigned int start, - unsigned int size, int must_exist); -extern void mem_pieces_coalesce(struct mem_pieces *mp); -extern void mem_pieces_sort(struct mem_pieces *mp); - -#ifdef __cplusplus -} -#endif - -#endif /* __MEM_PIECES_H__ */ diff --git a/arch/powerpc/mm/mmu_decl.h b/arch/powerpc/mm/mmu_decl.h index 540f3292b229..06fe8af3af55 100644 --- a/arch/powerpc/mm/mmu_decl.h +++ b/arch/powerpc/mm/mmu_decl.h @@ -36,6 +36,8 @@ extern unsigned long ioremap_base; extern unsigned long ioremap_bot; extern unsigned int rtas_data, rtas_size; +extern unsigned long __max_low_memory; +extern unsigned long __initial_memory_limit; extern unsigned long total_memory; extern unsigned long total_lowmem; extern int mem_init_done; diff --git a/arch/powerpc/mm/pgtable.c b/arch/powerpc/mm/pgtable.c index 81a3d7446d37..5792e533916f 100644 --- a/arch/powerpc/mm/pgtable.c +++ b/arch/powerpc/mm/pgtable.c @@ -190,8 +190,7 @@ __ioremap(phys_addr_t addr, unsigned long size, unsigned long flags) * Don't allow anybody to remap normal RAM that we're using. * mem_init() sets high_memory so only do the check after that. */ - if ( mem_init_done && (p < virt_to_phys(high_memory)) ) - { + if (mem_init_done && (p < virt_to_phys(high_memory))) { printk("__ioremap(): phys addr "PHYS_FMT" is RAM lr %p\n", p, __builtin_return_address(0)); return NULL; diff --git a/arch/powerpc/mm/ppc_mmu.c b/arch/powerpc/mm/ppc_mmu.c index 9a381ed5eb21..cef9e83cc7e9 100644 --- a/arch/powerpc/mm/ppc_mmu.c +++ b/arch/powerpc/mm/ppc_mmu.c @@ -32,9 +32,9 @@ #include #include #include +#include #include "mmu_decl.h" -#include "mem_pieces.h" PTE *Hash, *Hash_end; unsigned long Hash_size, Hash_mask; @@ -215,17 +215,6 @@ void __init MMU_init_hw(void) #define MIN_N_HPTEG 1024 /* min 64kB hash table */ #endif -#ifdef CONFIG_POWER4 - /* The hash table has already been allocated and initialized - in prom.c */ - n_hpteg = Hash_size >> LG_HPTEG_SIZE; - lg_n_hpteg = __ilog2(n_hpteg); - - /* Remove the hash table from the available memory */ - if (Hash) - reserve_phys_mem(__pa(Hash), Hash_size); - -#else /* CONFIG_POWER4 */ /* * Allow 1 HPTE (1/8 HPTEG) for each page of memory. * This is less than the recommended amount, but then @@ -245,10 +234,10 @@ void __init MMU_init_hw(void) * Find some memory for the hash table. */ if ( ppc_md.progress ) ppc_md.progress("hash:find piece", 0x322); - Hash = mem_pieces_find(Hash_size, Hash_size); + Hash = __va(lmb_alloc_base(Hash_size, Hash_size, + __initial_memory_limit)); cacheable_memzero(Hash, Hash_size); _SDR1 = __pa(Hash) | SDR1_LOW_BITS; -#endif /* CONFIG_POWER4 */ Hash_end = (PTE *) ((unsigned long)Hash + Hash_size); diff --git a/include/asm-ppc/page.h b/include/asm-ppc/page.h index 4789dc024240..fc44f7ca62d7 100644 --- a/include/asm-ppc/page.h +++ b/include/asm-ppc/page.h @@ -34,6 +34,17 @@ typedef unsigned long pte_basic_t; #define PTE_FMT "%.8lx" #endif +/* align addr on a size boundary - adjust address up/down if needed */ +#define _ALIGN_UP(addr,size) (((addr)+((size)-1))&(~((size)-1))) +#define _ALIGN_DOWN(addr,size) ((addr)&(~((size)-1))) + +/* align addr on a size boundary - adjust address up if needed */ +#define _ALIGN(addr,size) _ALIGN_UP(addr,size) + +/* to align the pointer to the (next) page boundary */ +#define PAGE_ALIGN(addr) _ALIGN(addr, PAGE_SIZE) + + #undef STRICT_MM_TYPECHECKS #ifdef STRICT_MM_TYPECHECKS @@ -76,13 +87,6 @@ typedef unsigned long pgprot_t; #endif - -/* align addr on a size boundary - adjust address up if needed -- Cort */ -#define _ALIGN(addr,size) (((addr)+(size)-1)&(~((size)-1))) - -/* to align the pointer to the (next) page boundary */ -#define PAGE_ALIGN(addr) (((addr)+PAGE_SIZE-1)&PAGE_MASK) - struct page; extern void clear_pages(void *page, int order); static inline void clear_page(void *page) { clear_pages(page, 0); }