[patch 2/2] m68k: Discontinuous memory support

From: Geert Uytterhoeven
Date: Mon May 28 2007 - 15:19:42 EST


From: Roman Zippel <zippel@xxxxxxxxxxxxxx>

Fix support for discontinuous memory

Signed-off-by: Roman Zippel <zippel@xxxxxxxxxxxxxx>
Signed-off-by: Geert Uytterhoeven <geert@xxxxxxxxxxxxxx>
---
arch/m68k/Kconfig | 13 +++
arch/m68k/kernel/module.c | 3
arch/m68k/kernel/setup.c | 37 ++---------
arch/m68k/mm/init.c | 119 +++++++++++++++++++++++-------------
arch/m68k/mm/memory.c | 73 ----------------------
arch/m68k/mm/motorola.c | 101 ++++++++++++++++++++----------
arch/m68k/sun3/config.c | 2
include/asm-m68k/mmzone.h | 9 ++
include/asm-m68k/module.h | 1
include/asm-m68k/motorola_pgtable.h | 10 +--
include/asm-m68k/page.h | 52 ++++++++++++---
include/asm-m68k/pgalloc.h | 3
include/asm-m68k/pgtable.h | 17 -----
include/asm-m68k/sun3_pgtable.h | 4 -
include/asm-m68k/virtconvert.h | 48 +++-----------
mm/page_alloc.c | 2
16 files changed, 246 insertions(+), 248 deletions(-)

--- a/arch/m68k/Kconfig
+++ b/arch/m68k/Kconfig
@@ -355,8 +355,9 @@ config RMW_INSNS
adventurous.

config SINGLE_MEMORY_CHUNK
- bool "Use one physical chunk of memory only"
- depends on ADVANCED && !SUN3
+ bool "Use one physical chunk of memory only" if ADVANCED && !SUN3
+ default y if SUN3
+ select NEED_MULTIPLE_NODES
help
Ignore all but the first contiguous chunk of physical memory for VM
purposes. This will save a few bytes kernel size and may speed up
@@ -377,6 +378,14 @@ config 060_WRITETHROUGH
is hardwired on. The 53c710 SCSI driver is known to suffer from
this problem.

+config ARCH_DISCONTIGMEM_ENABLE
+ def_bool !SINGLE_MEMORY_CHUNK
+
+config NODES_SHIFT
+ int
+ default "3"
+ depends on !SINGLE_MEMORY_CHUNK
+
source "mm/Kconfig"

endmenu
--- a/arch/m68k/kernel/module.c
+++ b/arch/m68k/kernel/module.c
@@ -149,6 +149,9 @@ void module_fixup(struct module *mod, st
case m68k_fixup_memoffset:
*(u32 *)fixup->addr = m68k_memoffset;
break;
+ case m68k_fixup_vnode_shift:
+ *(u16 *)fixup->addr += m68k_virt_to_node_shift;
+ break;
}
}
}
--- a/arch/m68k/kernel/setup.c
+++ b/arch/m68k/kernel/setup.c
@@ -60,14 +60,12 @@ extern unsigned long availmem;
int m68k_num_memory;
int m68k_realnum_memory;
EXPORT_SYMBOL(m68k_realnum_memory);
-#ifdef CONFIG_SINGLE_MEMORY_CHUNK
unsigned long m68k_memoffset;
EXPORT_SYMBOL(m68k_memoffset);
-#endif
struct mem_info m68k_memory[NUM_MEMINFO];
EXPORT_SYMBOL(m68k_memory);

-static struct mem_info m68k_ramdisk;
+struct mem_info m68k_ramdisk;

static char m68k_command_line[CL_SIZE];

@@ -208,9 +206,6 @@ static void __init m68k_parse_bootinfo(c
void __init setup_arch(char **cmdline_p)
{
extern int _etext, _edata, _end;
-#ifndef CONFIG_SUN3
- unsigned long endmem, startmem;
-#endif
int i;

/* The bootinfo is located right after the kernel bss */
@@ -320,30 +315,16 @@ void __init setup_arch(char **cmdline_p)
panic("No configuration setup");
}

-#ifndef CONFIG_SUN3
- startmem= m68k_memory[0].addr;
- endmem = startmem + m68k_memory[0].size;
- high_memory = (void *)PAGE_OFFSET;
- for (i = 0; i < m68k_num_memory; i++) {
- m68k_memory[i].size &= MASK_256K;
- if (m68k_memory[i].addr < startmem)
- startmem = m68k_memory[i].addr;
- if (m68k_memory[i].addr+m68k_memory[i].size > endmem)
- endmem = m68k_memory[i].addr+m68k_memory[i].size;
- high_memory += m68k_memory[i].size;
- }
-
- availmem += init_bootmem_node(NODE_DATA(0), availmem >> PAGE_SHIFT,
- startmem >> PAGE_SHIFT, endmem >> PAGE_SHIFT);
-
- for (i = 0; i < m68k_num_memory; i++)
- free_bootmem(m68k_memory[i].addr, m68k_memory[i].size);
-
- reserve_bootmem(m68k_memory[0].addr, availmem - m68k_memory[0].addr);
+ paging_init();

+#ifndef CONFIG_SUN3
+ for (i = 1; i < m68k_num_memory; i++)
+ free_bootmem_node(NODE_DATA(i), m68k_memory[i].addr,
+ m68k_memory[i].size);
#ifdef CONFIG_BLK_DEV_INITRD
if (m68k_ramdisk.size) {
- reserve_bootmem(m68k_ramdisk.addr, m68k_ramdisk.size);
+ reserve_bootmem_node(__virt_to_node(phys_to_virt(m68k_ramdisk.addr)),
+ m68k_ramdisk.addr, m68k_ramdisk.size);
initrd_start = (unsigned long)phys_to_virt(m68k_ramdisk.addr);
initrd_end = initrd_start + m68k_ramdisk.size;
printk("initrd: %08lx - %08lx\n", initrd_start, initrd_end);
@@ -362,8 +343,6 @@ void __init setup_arch(char **cmdline_p)

#endif /* !CONFIG_SUN3 */

- paging_init();
-
/* set ISA defs early as possible */
#if defined(CONFIG_ISA) && defined(MULTI_ISA)
#if defined(CONFIG_Q40)
--- a/arch/m68k/mm/init.c
+++ b/arch/m68k/mm/init.c
@@ -7,6 +7,7 @@
* to motorola.c and sun3mmu.c
*/

+#include <linux/module.h>
#include <linux/signal.h>
#include <linux/sched.h>
#include <linux/mm.h>
@@ -31,6 +32,37 @@

DEFINE_PER_CPU(struct mmu_gather, mmu_gathers);

+static bootmem_data_t __initdata bootmem_data[MAX_NUMNODES];
+
+pg_data_t pg_data_map[MAX_NUMNODES];
+EXPORT_SYMBOL(pg_data_map);
+
+int m68k_virt_to_node_shift;
+
+#ifndef CONFIG_SINGLE_MEMORY_CHUNK
+pg_data_t *pg_data_table[65];
+EXPORT_SYMBOL(pg_data_table);
+#endif
+
+void m68k_setup_node(int node)
+{
+#ifndef CONFIG_SINGLE_MEMORY_CHUNK
+ struct mem_info *info = m68k_memory + node;
+ int i, end;
+
+ i = (unsigned long)phys_to_virt(info->addr) >> __virt_to_node_shift();
+ end = (unsigned long)phys_to_virt(info->addr + info->size - 1) >> __virt_to_node_shift();
+ for (; i <= end; i++) {
+ if (pg_data_table[i])
+ printk("overlap at %u for chunk %u\n", i, node);
+ pg_data_table[i] = pg_data_map + node;
+ }
+#endif
+ pg_data_map[node].bdata = bootmem_data + node;
+ node_set_online(node);
+}
+
+
/*
* ZERO_PAGE is a special page that is used for zero-initialized
* data and COW.
@@ -40,52 +72,51 @@ void *empty_zero_page;

void show_mem(void)
{
- unsigned long i;
- int free = 0, total = 0, reserved = 0, shared = 0;
- int cached = 0;
-
- printk("\nMem-info:\n");
- show_free_areas();
- printk("Free swap: %6ldkB\n", nr_swap_pages<<(PAGE_SHIFT-10));
- 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;
- }
- printk("%d pages of RAM\n",total);
- printk("%d free pages\n",free);
- printk("%d reserved pages\n",reserved);
- printk("%d pages shared\n",shared);
- printk("%d pages swap cached\n",cached);
+ pg_data_t *pgdat;
+ int free = 0, total = 0, reserved = 0, shared = 0;
+ int cached = 0;
+ int i;
+
+ printk("\nMem-info:\n");
+ show_free_areas();
+ printk("Free swap: %6ldkB\n", nr_swap_pages<<(PAGE_SHIFT-10));
+ for_each_online_pgdat(pgdat) {
+ for (i = 0; i < pgdat->node_spanned_pages; i++) {
+ struct page *page = pgdat->node_mem_map + i;
+ total++;
+ if (PageReserved(page))
+ reserved++;
+ else if (PageSwapCache(page))
+ cached++;
+ else if (!page_count(page))
+ free++;
+ else
+ shared += page_count(page) - 1;
+ }
+ }
+ printk("%d pages of RAM\n",total);
+ printk("%d free pages\n",free);
+ printk("%d reserved pages\n",reserved);
+ printk("%d pages shared\n",shared);
+ printk("%d pages swap cached\n",cached);
}

extern void init_pointer_table(unsigned long ptable);

/* References to section boundaries */

-extern char _text, _etext, _edata, __bss_start, _end;
-extern char __init_begin, __init_end;
+extern char _text[], _etext[];
+extern char __init_begin[], __init_end[];

extern pmd_t *zero_pgtable;

void __init mem_init(void)
{
+ pg_data_t *pgdat;
int codepages = 0;
int datapages = 0;
int initpages = 0;
- unsigned long tmp;
-#ifndef CONFIG_SUN3
int i;
-#endif
-
- max_mapnr = num_physpages = (((unsigned long)high_memory - PAGE_OFFSET) >> PAGE_SHIFT);

#ifdef CONFIG_ATARI
if (MACH_IS_ATARI)
@@ -93,19 +124,25 @@ void __init mem_init(void)
#endif

/* this will put all memory onto the freelists */
- totalram_pages = free_all_bootmem();
-
- for (tmp = PAGE_OFFSET ; tmp < (unsigned long)high_memory; tmp += PAGE_SIZE) {
- if (PageReserved(virt_to_page(tmp))) {
- if (tmp >= (unsigned long)&_text
- && tmp < (unsigned long)&_etext)
+ totalram_pages = num_physpages = 0;
+ for_each_online_pgdat(pgdat) {
+ num_physpages += pgdat->node_present_pages;
+
+ totalram_pages += free_all_bootmem_node(pgdat);
+ for (i = 0; i < pgdat->node_spanned_pages; i++) {
+ struct page *page = pgdat->node_mem_map + i;
+ char *addr = page_to_virt(page);
+
+ if (!PageReserved(page))
+ continue;
+ if (addr >= _text &&
+ addr < _etext)
codepages++;
- else if (tmp >= (unsigned long) &__init_begin
- && tmp < (unsigned long) &__init_end)
+ else if (addr >= __init_begin &&
+ addr < __init_end)
initpages++;
else
datapages++;
- continue;
}
}

@@ -124,7 +161,7 @@ void __init mem_init(void)

printk("Memory: %luk/%luk available (%dk kernel code, %dk data, %dk init)\n",
(unsigned long)nr_free_pages() << (PAGE_SHIFT-10),
- max_mapnr << (PAGE_SHIFT-10),
+ totalram_pages << (PAGE_SHIFT-10),
codepages << (PAGE_SHIFT-10),
datapages << (PAGE_SHIFT-10),
initpages << (PAGE_SHIFT-10));
--- a/arch/m68k/mm/memory.c
+++ b/arch/m68k/mm/memory.c
@@ -127,67 +127,6 @@ int free_pointer_table (pmd_t *ptable)
return 0;
}

-#ifdef DEBUG_INVALID_PTOV
-int mm_inv_cnt = 5;
-#endif
-
-#ifndef CONFIG_SINGLE_MEMORY_CHUNK
-/*
- * The following two routines map from a physical address to a kernel
- * virtual address and vice versa.
- */
-unsigned long mm_vtop(unsigned long vaddr)
-{
- int i=0;
- unsigned long voff = (unsigned long)vaddr - PAGE_OFFSET;
-
- do {
- if (voff < m68k_memory[i].size) {
-#ifdef DEBUGPV
- printk ("VTOP(%p)=%lx\n", vaddr,
- m68k_memory[i].addr + voff);
-#endif
- return m68k_memory[i].addr + voff;
- }
- voff -= m68k_memory[i].size;
- } while (++i < m68k_num_memory);
-
- /* As a special case allow `__pa(high_memory)'. */
- if (voff == 0)
- return m68k_memory[i-1].addr + m68k_memory[i-1].size;
-
- return -1;
-}
-EXPORT_SYMBOL(mm_vtop);
-
-unsigned long mm_ptov (unsigned long paddr)
-{
- int i = 0;
- unsigned long poff, voff = PAGE_OFFSET;
-
- do {
- poff = paddr - m68k_memory[i].addr;
- if (poff < m68k_memory[i].size) {
-#ifdef DEBUGPV
- printk ("PTOV(%lx)=%lx\n", paddr, poff + voff);
-#endif
- return poff + voff;
- }
- voff += m68k_memory[i].size;
- } while (++i < m68k_num_memory);
-
-#ifdef DEBUG_INVALID_PTOV
- if (mm_inv_cnt > 0) {
- mm_inv_cnt--;
- printk("Invalid use of phys_to_virt(0x%lx) at 0x%p!\n",
- paddr, __builtin_return_address(0));
- }
-#endif
- return -1;
-}
-EXPORT_SYMBOL(mm_ptov);
-#endif
-
/* invalidate page in both caches */
static inline void clear040(unsigned long paddr)
{
@@ -354,15 +293,3 @@ void cache_push (unsigned long paddr, in
}
EXPORT_SYMBOL(cache_push);

-#ifndef CONFIG_SINGLE_MEMORY_CHUNK
-int mm_end_of_chunk (unsigned long addr, int len)
-{
- int i;
-
- for (i = 0; i < m68k_num_memory; i++)
- if (m68k_memory[i].addr + m68k_memory[i].size == addr + len)
- return 1;
- return 0;
-}
-EXPORT_SYMBOL(mm_end_of_chunk);
-#endif
--- a/arch/m68k/mm/motorola.c
+++ b/arch/m68k/mm/motorola.c
@@ -43,6 +43,11 @@ unsigned long mm_cachebits;
EXPORT_SYMBOL(mm_cachebits);
#endif

+/* size of memory already mapped in head.S */
+#define INIT_MAPPED_SIZE (4UL<<20)
+
+extern unsigned long availmem;
+
static pte_t * __init kernel_page_table(void)
{
pte_t *ptablep;
@@ -98,19 +103,20 @@ static pmd_t * __init kernel_ptr_table(v
return last_pgtable;
}

-static unsigned long __init
-map_chunk (unsigned long addr, long size)
+static void __init map_node(int node)
{
#define PTRTREESIZE (256*1024)
#define ROOTTREESIZE (32*1024*1024)
- static unsigned long virtaddr = PAGE_OFFSET;
- unsigned long physaddr;
+ unsigned long physaddr, virtaddr, size;
pgd_t *pgd_dir;
pmd_t *pmd_dir;
pte_t *pte_dir;

- physaddr = (addr | m68k_supervisor_cachemode |
- _PAGE_PRESENT | _PAGE_ACCESSED | _PAGE_DIRTY);
+ size = m68k_memory[node].size;
+ physaddr = m68k_memory[node].addr;
+ virtaddr = (unsigned long)phys_to_virt(physaddr);
+ physaddr |= m68k_supervisor_cachemode |
+ _PAGE_PRESENT | _PAGE_ACCESSED | _PAGE_DIRTY;
if (CPU_IS_040_OR_060)
physaddr |= _PAGE_GLOBAL040;

@@ -190,8 +196,6 @@ map_chunk (unsigned long addr, long size
#ifdef DEBUG
printk("\n");
#endif
-
- return virtaddr;
}

/*
@@ -200,15 +204,16 @@ map_chunk (unsigned long addr, long size
*/
void __init paging_init(void)
{
- int chunk;
- unsigned long mem_avail = 0;
unsigned long zones_size[MAX_NR_ZONES] = { 0, };
+ unsigned long min_addr, max_addr;
+ unsigned long addr, size, end;
+ int i;

#ifdef DEBUG
{
extern unsigned long availmem;
- printk ("start of paging_init (%p, %lx, %lx, %lx)\n",
- kernel_pg_dir, availmem, start_mem, end_mem);
+ printk ("start of paging_init (%p, %lx)\n",
+ kernel_pg_dir, availmem);
}
#endif

@@ -222,27 +227,62 @@ void __init paging_init(void)
pgprot_val(protection_map[i]) |= _PAGE_CACHE040;
}

+ min_addr = m68k_memory[0].addr;
+ max_addr = min_addr + m68k_memory[0].size;
+ for (i = 1; i < m68k_num_memory;) {
+ if (m68k_memory[i].addr < min_addr) {
+ printk("Ignoring memory chunk at 0x%lx:0x%lx before the first chunk\n",
+ m68k_memory[i].addr, m68k_memory[i].size);
+ printk("Fix your bootloader or use a memfile to make use of this area!\n");
+ m68k_num_memory--;
+ memmove(m68k_memory + i, m68k_memory + i + 1,
+ (m68k_num_memory - i) * sizeof(struct mem_info));
+ continue;
+ }
+ addr = m68k_memory[i].addr + m68k_memory[i].size;
+ if (addr > max_addr)
+ max_addr = addr;
+ i++;
+ }
+ m68k_memoffset = min_addr - PAGE_OFFSET;
+ m68k_virt_to_node_shift = fls(max_addr - min_addr - 1) - 6;
+
module_fixup(NULL, __start_fixup, __stop_fixup);
flush_icache();

+ high_memory = phys_to_virt(max_addr);
+
+ min_low_pfn = availmem >> PAGE_SHIFT;
+ max_low_pfn = max_addr >> PAGE_SHIFT;
+
+ for (i = 0; i < m68k_num_memory; i++) {
+ addr = m68k_memory[i].addr;
+ end = addr + m68k_memory[i].size;
+ m68k_setup_node(i);
+ availmem = PAGE_ALIGN(availmem);
+ availmem += init_bootmem_node(NODE_DATA(i),
+ availmem >> PAGE_SHIFT,
+ addr >> PAGE_SHIFT,
+ end >> PAGE_SHIFT);
+ }
+
/*
* Map the physical memory available into the kernel virtual
- * address space. It may allocate some memory for page
- * tables and thus modify availmem.
+ * address space. First initialize the bootmem allocator with
+ * the memory we already mapped, so map_node() has something
+ * to allocate.
*/
+ addr = m68k_memory[0].addr;
+ size = m68k_memory[0].size;
+ free_bootmem_node(NODE_DATA(0), availmem, min(INIT_MAPPED_SIZE, size) - (availmem - addr));
+ map_node(0);
+ if (size > INIT_MAPPED_SIZE)
+ free_bootmem_node(NODE_DATA(0), addr + INIT_MAPPED_SIZE, size - INIT_MAPPED_SIZE);

- for (chunk = 0; chunk < m68k_num_memory; chunk++) {
- mem_avail = map_chunk (m68k_memory[chunk].addr,
- m68k_memory[chunk].size);
-
- }
+ for (i = 1; i < m68k_num_memory; i++)
+ map_node(i);

flush_tlb_all();
-#ifdef DEBUG
- printk ("memory available is %ldKB\n", mem_avail >> 10);
- printk ("start_mem is %#lx\nvirtual_end is %#lx\n",
- start_mem, end_mem);
-#endif

/*
* initialize the bad page table and bad page to point
@@ -259,14 +299,11 @@ void __init paging_init(void)
#ifdef DEBUG
printk ("before free_area_init\n");
#endif
- zones_size[ZONE_DMA] = (mach_max_dma_address < (unsigned long)high_memory ?
- (mach_max_dma_address+1) : (unsigned long)high_memory);
- zones_size[ZONE_NORMAL] = (unsigned long)high_memory - zones_size[0];
-
- zones_size[ZONE_DMA] = (zones_size[ZONE_DMA] - PAGE_OFFSET) >> PAGE_SHIFT;
- zones_size[ZONE_NORMAL] >>= PAGE_SHIFT;
-
- free_area_init(zones_size);
+ for (i = 0; i < m68k_num_memory; i++) {
+ zones_size[ZONE_DMA] = m68k_memory[i].size >> PAGE_SHIFT;
+ free_area_init_node(i, pg_data_map + i, zones_size,
+ m68k_memory[i].addr >> PAGE_SHIFT, NULL);
+ }
}

extern char __init_begin, __init_end;
--- a/arch/m68k/sun3/config.c
+++ b/arch/m68k/sun3/config.c
@@ -21,6 +21,7 @@
#include <asm/contregs.h>
#include <asm/movs.h>
#include <asm/pgtable.h>
+#include <asm/pgalloc.h>
#include <asm/sun3-head.h>
#include <asm/sun3mmu.h>
#include <asm/rtc.h>
@@ -127,6 +128,7 @@ void __init sun3_bootmem_alloc(unsigned
high_memory = (void *)memory_end;
availmem = memory_start;

+ m68k_setup_node(0);
availmem += init_bootmem_node(NODE_DATA(0), start_page, 0, num_pages);
availmem = (availmem + (PAGE_SIZE-1)) & PAGE_MASK;

--- /dev/null
+++ b/include/asm-m68k/mmzone.h
@@ -0,0 +1,9 @@
+#ifndef _ASM_M68K_MMZONE_H_
+#define _ASM_M68K_MMZONE_H_
+
+extern pg_data_t pg_data_map[];
+
+#define NODE_DATA(nid) (&pg_data_map[nid])
+#define NODE_MEM_MAP(nid) (NODE_DATA(nid)->node_mem_map)
+
+#endif /* _ASM_M68K_MMZONE_H_ */
--- a/include/asm-m68k/module.h
+++ b/include/asm-m68k/module.h
@@ -17,6 +17,7 @@ struct mod_arch_specific {

enum m68k_fixup_type {
m68k_fixup_memoffset,
+ m68k_fixup_vnode_shift,
};

struct m68k_fixup_info {
--- a/include/asm-m68k/motorola_pgtable.h
+++ b/include/asm-m68k/motorola_pgtable.h
@@ -130,7 +130,7 @@ static inline void pgd_set(pgd_t *pgdp,
#define pte_present(pte) (pte_val(pte) & (_PAGE_PRESENT | _PAGE_PROTNONE))
#define pte_clear(mm,addr,ptep) ({ pte_val(*(ptep)) = 0; })

-#define pte_page(pte) (mem_map + ((unsigned long)(__va(pte_val(pte)) - PAGE_OFFSET) >> PAGE_SHIFT))
+#define pte_page(pte) virt_to_page(__va(pte_val(pte)))
#define pte_pfn(pte) (pte_val(pte) >> PAGE_SHIFT)
#define pfn_pte(pfn, prot) __pte(((pfn) << PAGE_SHIFT) | pgprot_val(prot))

@@ -143,7 +143,7 @@ static inline void pgd_set(pgd_t *pgdp,
while (--__i >= 0) \
*__ptr++ = 0; \
})
-#define pmd_page(pmd) (mem_map + ((unsigned long)(__va(pmd_val(pmd)) - PAGE_OFFSET) >> PAGE_SHIFT))
+#define pmd_page(pmd) virt_to_page(__va(pmd_val(pmd)))


#define pgd_none(pgd) (!pgd_val(pgd))
@@ -223,10 +223,10 @@ static inline pte_t *pte_offset_kernel(p
return (pte_t *)__pmd_page(*pmdp) + ((address >> PAGE_SHIFT) & (PTRS_PER_PTE - 1));
}

-#define pte_offset_map(pmdp,address) ((pte_t *)kmap(pmd_page(*pmdp)) + ((address >> PAGE_SHIFT) & (PTRS_PER_PTE - 1)))
+#define pte_offset_map(pmdp,address) ((pte_t *)__pmd_page(*pmdp) + (((address) >> PAGE_SHIFT) & (PTRS_PER_PTE - 1)))
#define pte_offset_map_nested(pmdp, address) pte_offset_map(pmdp, address)
-#define pte_unmap(pte) kunmap(pte)
-#define pte_unmap_nested(pte) kunmap(pte)
+#define pte_unmap(pte) ((void)0)
+#define pte_unmap_nested(pte) ((void)0)

/*
* Allocate and free page tables. The xxx_kernel() versions are
--- a/include/asm-m68k/page.h
+++ b/include/asm-m68k/page.h
@@ -121,7 +121,6 @@ extern unsigned long m68k_memoffset;
#ifndef CONFIG_SUN3

#define WANT_PAGE_VIRTUAL
-#ifdef CONFIG_SINGLE_MEMORY_CHUNK

static inline unsigned long ___pa(void *vaddr)
{
@@ -133,7 +132,7 @@ static inline unsigned long ___pa(void *
: "0" (vaddr), "i" (m68k_fixup_memoffset));
return paddr;
}
-#define __pa(vaddr) ___pa((void *)(vaddr))
+#define __pa(vaddr) ___pa((void *)(vaddr))
static inline void *__va(unsigned long paddr)
{
void *vaddr;
@@ -145,11 +144,6 @@ static inline void *__va(unsigned long p
return vaddr;
}

-#else
-#define __pa(vaddr) virt_to_phys((void *)(vaddr))
-#define __va(paddr) phys_to_virt((unsigned long)(paddr))
-#endif
-
#else /* !CONFIG_SUN3 */
/* This #define is a horrible hack to suppress lots of warnings. --m */
#define __pa(x) ___pa((unsigned long)(x))
@@ -184,11 +178,47 @@ static inline void *__va(unsigned long x
#define virt_to_pfn(kaddr) (__pa(kaddr) >> PAGE_SHIFT)
#define pfn_to_virt(pfn) __va((pfn) << PAGE_SHIFT)

-#define virt_to_page(kaddr) (mem_map + (((unsigned long)(kaddr)-PAGE_OFFSET) >> PAGE_SHIFT))
-#define page_to_virt(page) ((((page) - mem_map) << PAGE_SHIFT) + PAGE_OFFSET)
+extern int m68k_virt_to_node_shift;
+
+#ifdef CONFIG_SINGLE_MEMORY_CHUNK
+#define __virt_to_node(addr) (&pg_data_map[0])
+#else
+extern struct pglist_data *pg_data_table[];
+
+static inline __attribute_const__ int __virt_to_node_shift(void)
+{
+ int shift;
+
+ asm (
+ "1: moveq #0,%0\n"
+ m68k_fixup(%c1, 1b)
+ : "=d" (shift)
+ : "i" (m68k_fixup_vnode_shift));
+ return shift;
+}
+
+#define __virt_to_node(addr) (pg_data_table[(unsigned long)(addr) >> __virt_to_node_shift()])
+#endif

-#define pfn_to_page(pfn) virt_to_page(pfn_to_virt(pfn))
-#define page_to_pfn(page) virt_to_pfn(page_to_virt(page))
+#define virt_to_page(addr) ({ \
+ pfn_to_page(virt_to_pfn(addr)); \
+})
+#define page_to_virt(page) ({ \
+ pfn_to_virt(page_to_pfn(page)); \
+})
+
+#define pfn_to_page(pfn) ({ \
+ unsigned long __pfn = (pfn); \
+ struct pglist_data *pgdat; \
+ pgdat = __virt_to_node((unsigned long)pfn_to_virt(__pfn)); \
+ pgdat->node_mem_map + (__pfn - pgdat->node_start_pfn); \
+})
+#define page_to_pfn(_page) ({ \
+ struct page *__p = (_page); \
+ struct pglist_data *pgdat; \
+ pgdat = &pg_data_map[page_to_nid(__p)]; \
+ ((__p) - pgdat->node_mem_map) + pgdat->node_start_pfn; \
+})

#define virt_addr_valid(kaddr) ((void *)(kaddr) >= (void *)PAGE_OFFSET && (void *)(kaddr) < high_memory)
#define pfn_valid(pfn) virt_addr_valid(pfn_to_virt(pfn))
--- a/include/asm-m68k/pgalloc.h
+++ b/include/asm-m68k/pgalloc.h
@@ -8,11 +8,12 @@
#include <asm/virtconvert.h>


-
#ifdef CONFIG_SUN3
#include <asm/sun3_pgalloc.h>
#else
#include <asm/motorola_pgalloc.h>
#endif

+extern void m68k_setup_node(int node);
+
#endif /* M68K_PGALLOC_H */
--- a/include/asm-m68k/pgtable.h
+++ b/include/asm-m68k/pgtable.h
@@ -107,22 +107,7 @@ extern void *empty_zero_page;
/* 64-bit machines, beware! SRB. */
#define SIZEOF_PTR_LOG2 2

-/*
- * Check if the addr/len goes up to the end of a physical
- * memory chunk. Used for DMA functions.
- */
-#ifdef CONFIG_SINGLE_MEMORY_CHUNK
-/*
- * It makes no sense to consider whether we cross a memory boundary if
- * we support just one physical chunk of memory.
- */
-static inline int mm_end_of_chunk(unsigned long addr, int len)
-{
- return 0;
-}
-#else
-int mm_end_of_chunk (unsigned long addr, int len);
-#endif
+#define mm_end_of_chunk(addr, len) 0

extern void kernel_set_cachemode(void *addr, unsigned long size, int cmode);

--- a/include/asm-m68k/sun3_pgtable.h
+++ b/include/asm-m68k/sun3_pgtable.h
@@ -132,8 +132,8 @@ static inline void pte_clear (struct mm_
#define pfn_pte(pfn, pgprot) \
({ pte_t __pte; pte_val(__pte) = pfn | pgprot_val(pgprot); __pte; })

-#define pte_page(pte) (mem_map+((__pte_page(pte) - PAGE_OFFSET) >> PAGE_SHIFT))
-#define pmd_page(pmd) (mem_map+((__pmd_page(pmd) - PAGE_OFFSET) >> PAGE_SHIFT))
+#define pte_page(pte) virt_to_page(__pte_page(pte))
+#define pmd_page(pmd) virt_to_page(__pmd_page(pmd))


static inline int pmd_none2 (pmd_t *pmd) { return !pmd_val (*pmd); }
--- a/include/asm-m68k/virtconvert.h
+++ b/include/asm-m68k/virtconvert.h
@@ -8,56 +8,34 @@
#ifdef __KERNEL__

#include <linux/compiler.h>
+#include <linux/mmzone.h>
#include <asm/setup.h>
#include <asm/page.h>

-#ifdef CONFIG_AMIGA
-#include <asm/amigahw.h>
-#endif
-
/*
* Change virtual addresses to physical addresses and vv.
*/
-#ifndef CONFIG_SUN3
-extern unsigned long mm_vtop(unsigned long addr) __attribute_const__;
-extern unsigned long mm_ptov(unsigned long addr) __attribute_const__;
-#else
-static inline unsigned long mm_vtop(unsigned long vaddr)
-{
- return __pa(vaddr);
-}
-
-static inline unsigned long mm_ptov(unsigned long paddr)
-{
- return (unsigned long)__va(paddr);
-}
-#endif
-
-#ifdef CONFIG_SINGLE_MEMORY_CHUNK
-static inline unsigned long virt_to_phys(void *vaddr)
-{
- return (unsigned long)vaddr - PAGE_OFFSET + m68k_memory[0].addr;
-}
-
-static inline void * phys_to_virt(unsigned long paddr)
-{
- return (void *)(paddr - m68k_memory[0].addr + PAGE_OFFSET);
-}
-#else
static inline unsigned long virt_to_phys(void *address)
{
- return mm_vtop((unsigned long)address);
+ return __pa(address);
}

static inline void *phys_to_virt(unsigned long address)
{
- return (void *) mm_ptov(address);
+ return __va(address);
}
-#endif

/* Permanent address of a page. */
-#define __page_address(page) (PAGE_OFFSET + (((page) - mem_map) << PAGE_SHIFT))
-#define page_to_phys(page) virt_to_phys((void *)__page_address(page))
+#ifdef CONFIG_SINGLE_MEMORY_CHUNK
+#define page_to_phys(page) \
+ __pa(PAGE_OFFSET + (((page) - pg_data_map[0].node_mem_map) << PAGE_SHIFT))
+#else
+#define page_to_phys(page) ({ \
+ struct pglist_data *pgdat; \
+ pgdat = pg_data_table[page_to_nid(page)]; \
+ page_to_pfn(page) << PAGE_SHIFT; \
+})
+#endif

/*
* IO bus memory addresses are 1:1 with the physical address,
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -2689,7 +2689,7 @@ static void __init_refok alloc_node_mem_
map = alloc_bootmem_node(pgdat, size);
pgdat->node_mem_map = map + (pgdat->node_start_pfn - start);
}
-#ifdef CONFIG_FLATMEM
+#ifndef CONFIG_NEED_MULTIPLE_NODES
/*
* With no DISCONTIG, the global mem_map is just set as node 0's
*/

--
Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/