2005-04-16 18:20:36 -04:00
|
|
|
/*
|
2008-07-24 00:28:00 -04:00
|
|
|
* bootmem - A boot-time physical memory allocator and configurator
|
2005-04-16 18:20:36 -04:00
|
|
|
*
|
|
|
|
* Copyright (C) 1999 Ingo Molnar
|
2008-07-24 00:28:00 -04:00
|
|
|
* 1999 Kanoj Sarcar, SGI
|
|
|
|
* 2008 Johannes Weiner
|
2005-04-16 18:20:36 -04:00
|
|
|
*
|
2008-07-24 00:28:00 -04:00
|
|
|
* Access to this subsystem has to be serialized externally (which is true
|
|
|
|
* for the boot process anyway).
|
2005-04-16 18:20:36 -04:00
|
|
|
*/
|
|
|
|
#include <linux/init.h>
|
2006-09-26 02:31:07 -04:00
|
|
|
#include <linux/pfn.h>
|
2005-04-16 18:20:36 -04:00
|
|
|
#include <linux/bootmem.h>
|
|
|
|
#include <linux/module.h>
|
2006-09-26 02:31:06 -04:00
|
|
|
|
|
|
|
#include <asm/bug.h>
|
2005-04-16 18:20:36 -04:00
|
|
|
#include <asm/io.h>
|
2006-09-26 02:31:33 -04:00
|
|
|
#include <asm/processor.h>
|
2006-09-26 02:31:06 -04:00
|
|
|
|
2005-04-16 18:20:36 -04:00
|
|
|
#include "internal.h"
|
|
|
|
|
|
|
|
unsigned long max_low_pfn;
|
|
|
|
unsigned long min_low_pfn;
|
|
|
|
unsigned long max_pfn;
|
|
|
|
|
2005-06-25 17:58:18 -04:00
|
|
|
#ifdef CONFIG_CRASH_DUMP
|
|
|
|
/*
|
|
|
|
* If we have booted due to a crash, max_pfn will be a very low value. We need
|
|
|
|
* to know the amount of memory that the previous kernel used.
|
|
|
|
*/
|
|
|
|
unsigned long saved_max_pfn;
|
|
|
|
#endif
|
|
|
|
|
2008-07-24 00:26:55 -04:00
|
|
|
bootmem_data_t bootmem_node_data[MAX_NUMNODES] __initdata;
|
|
|
|
|
2008-07-24 00:28:03 -04:00
|
|
|
static struct list_head bdata_list __initdata = LIST_HEAD_INIT(bdata_list);
|
|
|
|
|
2008-07-24 00:28:02 -04:00
|
|
|
static int bootmem_debug;
|
|
|
|
|
|
|
|
static int __init bootmem_debug_setup(char *buf)
|
|
|
|
{
|
|
|
|
bootmem_debug = 1;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
early_param("bootmem_debug", bootmem_debug_setup);
|
|
|
|
|
|
|
|
#define bdebug(fmt, args...) ({ \
|
|
|
|
if (unlikely(bootmem_debug)) \
|
|
|
|
printk(KERN_INFO \
|
|
|
|
"bootmem::%s " fmt, \
|
2008-10-16 01:01:25 -04:00
|
|
|
__func__, ## args); \
|
2008-07-24 00:28:02 -04:00
|
|
|
})
|
|
|
|
|
2008-07-24 00:28:02 -04:00
|
|
|
static unsigned long __init bootmap_bytes(unsigned long pages)
|
2008-07-24 00:28:00 -04:00
|
|
|
{
|
2008-07-24 00:28:02 -04:00
|
|
|
unsigned long bytes = (pages + 7) / 8;
|
2008-07-24 00:28:00 -04:00
|
|
|
|
2008-07-24 00:28:02 -04:00
|
|
|
return ALIGN(bytes, sizeof(long));
|
2008-07-24 00:28:00 -04:00
|
|
|
}
|
|
|
|
|
2008-07-24 00:28:01 -04:00
|
|
|
/**
|
|
|
|
* bootmem_bootmap_pages - calculate bitmap size in pages
|
|
|
|
* @pages: number of pages the bitmap has to represent
|
|
|
|
*/
|
2006-09-26 02:31:08 -04:00
|
|
|
unsigned long __init bootmem_bootmap_pages(unsigned long pages)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
2008-07-24 00:28:02 -04:00
|
|
|
unsigned long bytes = bootmap_bytes(pages);
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2008-07-24 00:28:02 -04:00
|
|
|
return PAGE_ALIGN(bytes) >> PAGE_SHIFT;
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
2006-09-26 02:31:08 -04:00
|
|
|
|
2006-03-27 04:15:58 -05:00
|
|
|
/*
|
|
|
|
* link bdata in order
|
|
|
|
*/
|
2006-09-26 02:31:04 -04:00
|
|
|
static void __init link_bootmem(bootmem_data_t *bdata)
|
2006-03-27 04:15:58 -05:00
|
|
|
{
|
2008-07-24 00:28:03 -04:00
|
|
|
struct list_head *iter;
|
2006-09-26 02:31:08 -04:00
|
|
|
|
2008-07-24 00:28:03 -04:00
|
|
|
list_for_each(iter, &bdata_list) {
|
|
|
|
bootmem_data_t *ent;
|
|
|
|
|
|
|
|
ent = list_entry(iter, bootmem_data_t, list);
|
2008-07-24 00:28:09 -04:00
|
|
|
if (bdata->node_min_pfn < ent->node_min_pfn)
|
2008-07-24 00:28:03 -04:00
|
|
|
break;
|
2006-03-27 04:15:58 -05:00
|
|
|
}
|
2008-07-24 00:28:03 -04:00
|
|
|
list_add_tail(&bdata->list, iter);
|
2006-03-27 04:15:58 -05:00
|
|
|
}
|
|
|
|
|
2005-04-16 18:20:36 -04:00
|
|
|
/*
|
|
|
|
* Called once to set up the allocator itself.
|
|
|
|
*/
|
2008-07-24 00:26:59 -04:00
|
|
|
static unsigned long __init init_bootmem_core(bootmem_data_t *bdata,
|
2005-04-16 18:20:36 -04:00
|
|
|
unsigned long mapstart, unsigned long start, unsigned long end)
|
|
|
|
{
|
2006-09-26 02:31:07 -04:00
|
|
|
unsigned long mapsize;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2008-07-24 00:26:52 -04:00
|
|
|
mminit_validate_memmodel_limits(&start, &end);
|
2006-09-26 02:31:07 -04:00
|
|
|
bdata->node_bootmem_map = phys_to_virt(PFN_PHYS(mapstart));
|
2008-07-24 00:28:09 -04:00
|
|
|
bdata->node_min_pfn = start;
|
2005-04-16 18:20:36 -04:00
|
|
|
bdata->node_low_pfn = end;
|
2006-03-27 04:15:58 -05:00
|
|
|
link_bootmem(bdata);
|
2005-04-16 18:20:36 -04:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Initially all pages are reserved - setup_arch() has to
|
|
|
|
* register free RAM areas explicitly.
|
|
|
|
*/
|
2008-07-24 00:28:02 -04:00
|
|
|
mapsize = bootmap_bytes(end - start);
|
2005-04-16 18:20:36 -04:00
|
|
|
memset(bdata->node_bootmem_map, 0xff, mapsize);
|
|
|
|
|
2008-07-24 00:28:02 -04:00
|
|
|
bdebug("nid=%td start=%lx map=%lx end=%lx mapsize=%lx\n",
|
|
|
|
bdata - bootmem_node_data, start, mapstart, end, mapsize);
|
|
|
|
|
2005-04-16 18:20:36 -04:00
|
|
|
return mapsize;
|
|
|
|
}
|
|
|
|
|
2008-07-24 00:28:01 -04:00
|
|
|
/**
|
|
|
|
* init_bootmem_node - register a node as boot memory
|
|
|
|
* @pgdat: node to register
|
|
|
|
* @freepfn: pfn where the bitmap for this node is to be placed
|
|
|
|
* @startpfn: first pfn on the node
|
|
|
|
* @endpfn: first pfn after the node
|
|
|
|
*
|
|
|
|
* Returns the number of bytes needed to hold the bitmap for this node.
|
|
|
|
*/
|
2008-07-24 00:28:00 -04:00
|
|
|
unsigned long __init init_bootmem_node(pg_data_t *pgdat, unsigned long freepfn,
|
|
|
|
unsigned long startpfn, unsigned long endpfn)
|
|
|
|
{
|
|
|
|
return init_bootmem_core(pgdat->bdata, freepfn, startpfn, endpfn);
|
|
|
|
}
|
|
|
|
|
2008-07-24 00:28:01 -04:00
|
|
|
/**
|
|
|
|
* init_bootmem - register boot memory
|
|
|
|
* @start: pfn where the bitmap is to be placed
|
|
|
|
* @pages: number of available physical pages
|
|
|
|
*
|
|
|
|
* Returns the number of bytes needed to hold the bitmap.
|
|
|
|
*/
|
2008-07-24 00:28:00 -04:00
|
|
|
unsigned long __init init_bootmem(unsigned long start, unsigned long pages)
|
|
|
|
{
|
|
|
|
max_low_pfn = pages;
|
|
|
|
min_low_pfn = start;
|
|
|
|
return init_bootmem_core(NODE_DATA(0)->bdata, start, 0, pages);
|
|
|
|
}
|
|
|
|
|
|
|
|
static unsigned long __init free_all_bootmem_core(bootmem_data_t *bdata)
|
|
|
|
{
|
2008-07-24 00:28:03 -04:00
|
|
|
int aligned;
|
2008-07-24 00:28:00 -04:00
|
|
|
struct page *page;
|
2008-07-24 00:28:03 -04:00
|
|
|
unsigned long start, end, pages, count = 0;
|
|
|
|
|
|
|
|
if (!bdata->node_bootmem_map)
|
|
|
|
return 0;
|
|
|
|
|
2008-07-24 00:28:09 -04:00
|
|
|
start = bdata->node_min_pfn;
|
2008-07-24 00:28:03 -04:00
|
|
|
end = bdata->node_low_pfn;
|
|
|
|
|
2008-07-24 00:28:00 -04:00
|
|
|
/*
|
2008-07-24 00:28:03 -04:00
|
|
|
* If the start is aligned to the machines wordsize, we might
|
|
|
|
* be able to free pages in bulks of that order.
|
2008-07-24 00:28:00 -04:00
|
|
|
*/
|
2008-07-24 00:28:03 -04:00
|
|
|
aligned = !(start & (BITS_PER_LONG - 1));
|
2008-07-24 00:28:00 -04:00
|
|
|
|
2008-07-24 00:28:03 -04:00
|
|
|
bdebug("nid=%td start=%lx end=%lx aligned=%d\n",
|
|
|
|
bdata - bootmem_node_data, start, end, aligned);
|
2008-07-24 00:28:00 -04:00
|
|
|
|
2008-07-24 00:28:03 -04:00
|
|
|
while (start < end) {
|
|
|
|
unsigned long *map, idx, vec;
|
2008-07-24 00:28:00 -04:00
|
|
|
|
2008-07-24 00:28:03 -04:00
|
|
|
map = bdata->node_bootmem_map;
|
2008-07-24 00:28:09 -04:00
|
|
|
idx = start - bdata->node_min_pfn;
|
2008-07-24 00:28:03 -04:00
|
|
|
vec = ~map[idx / BITS_PER_LONG];
|
|
|
|
|
|
|
|
if (aligned && vec == ~0UL && start + BITS_PER_LONG < end) {
|
|
|
|
int order = ilog2(BITS_PER_LONG);
|
|
|
|
|
|
|
|
__free_pages_bootmem(pfn_to_page(start), order);
|
2008-07-24 00:28:00 -04:00
|
|
|
count += BITS_PER_LONG;
|
2008-07-24 00:28:03 -04:00
|
|
|
} else {
|
|
|
|
unsigned long off = 0;
|
|
|
|
|
|
|
|
while (vec && off < BITS_PER_LONG) {
|
|
|
|
if (vec & 1) {
|
|
|
|
page = pfn_to_page(start + off);
|
2008-07-24 00:28:00 -04:00
|
|
|
__free_pages_bootmem(page, 0);
|
2008-07-24 00:28:03 -04:00
|
|
|
count++;
|
2008-07-24 00:28:00 -04:00
|
|
|
}
|
2008-07-24 00:28:03 -04:00
|
|
|
vec >>= 1;
|
|
|
|
off++;
|
2008-07-24 00:28:00 -04:00
|
|
|
}
|
|
|
|
}
|
2008-07-24 00:28:03 -04:00
|
|
|
start += BITS_PER_LONG;
|
2008-07-24 00:28:00 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
page = virt_to_page(bdata->node_bootmem_map);
|
2008-07-24 00:28:09 -04:00
|
|
|
pages = bdata->node_low_pfn - bdata->node_min_pfn;
|
2008-07-24 00:28:03 -04:00
|
|
|
pages = bootmem_bootmap_pages(pages);
|
|
|
|
count += pages;
|
|
|
|
while (pages--)
|
|
|
|
__free_pages_bootmem(page++, 0);
|
2008-07-24 00:28:00 -04:00
|
|
|
|
2008-07-24 00:28:02 -04:00
|
|
|
bdebug("nid=%td released=%lx\n", bdata - bootmem_node_data, count);
|
|
|
|
|
2008-07-24 00:28:00 -04:00
|
|
|
return count;
|
|
|
|
}
|
|
|
|
|
2008-07-24 00:28:01 -04:00
|
|
|
/**
|
|
|
|
* free_all_bootmem_node - release a node's free pages to the buddy allocator
|
|
|
|
* @pgdat: node to be released
|
|
|
|
*
|
|
|
|
* Returns the number of pages actually released.
|
|
|
|
*/
|
2008-07-24 00:28:00 -04:00
|
|
|
unsigned long __init free_all_bootmem_node(pg_data_t *pgdat)
|
|
|
|
{
|
|
|
|
register_page_bootmem_info_node(pgdat);
|
|
|
|
return free_all_bootmem_core(pgdat->bdata);
|
|
|
|
}
|
|
|
|
|
2008-07-24 00:28:01 -04:00
|
|
|
/**
|
|
|
|
* free_all_bootmem - release free pages to the buddy allocator
|
|
|
|
*
|
|
|
|
* Returns the number of pages actually released.
|
|
|
|
*/
|
2008-07-24 00:28:00 -04:00
|
|
|
unsigned long __init free_all_bootmem(void)
|
|
|
|
{
|
|
|
|
return free_all_bootmem_core(NODE_DATA(0)->bdata);
|
|
|
|
}
|
|
|
|
|
2008-07-24 00:28:05 -04:00
|
|
|
static void __init __free(bootmem_data_t *bdata,
|
|
|
|
unsigned long sidx, unsigned long eidx)
|
|
|
|
{
|
|
|
|
unsigned long idx;
|
|
|
|
|
|
|
|
bdebug("nid=%td start=%lx end=%lx\n", bdata - bootmem_node_data,
|
2008-07-24 00:28:09 -04:00
|
|
|
sidx + bdata->node_min_pfn,
|
|
|
|
eidx + bdata->node_min_pfn);
|
2008-07-24 00:28:05 -04:00
|
|
|
|
2008-07-24 00:28:06 -04:00
|
|
|
if (bdata->hint_idx > sidx)
|
|
|
|
bdata->hint_idx = sidx;
|
|
|
|
|
2008-07-24 00:28:05 -04:00
|
|
|
for (idx = sidx; idx < eidx; idx++)
|
|
|
|
if (!test_and_clear_bit(idx, bdata->node_bootmem_map))
|
|
|
|
BUG();
|
|
|
|
}
|
|
|
|
|
|
|
|
static int __init __reserve(bootmem_data_t *bdata, unsigned long sidx,
|
|
|
|
unsigned long eidx, int flags)
|
|
|
|
{
|
|
|
|
unsigned long idx;
|
|
|
|
int exclusive = flags & BOOTMEM_EXCLUSIVE;
|
|
|
|
|
|
|
|
bdebug("nid=%td start=%lx end=%lx flags=%x\n",
|
|
|
|
bdata - bootmem_node_data,
|
2008-07-24 00:28:09 -04:00
|
|
|
sidx + bdata->node_min_pfn,
|
|
|
|
eidx + bdata->node_min_pfn,
|
2008-07-24 00:28:05 -04:00
|
|
|
flags);
|
|
|
|
|
|
|
|
for (idx = sidx; idx < eidx; idx++)
|
|
|
|
if (test_and_set_bit(idx, bdata->node_bootmem_map)) {
|
|
|
|
if (exclusive) {
|
|
|
|
__free(bdata, sidx, idx);
|
|
|
|
return -EBUSY;
|
|
|
|
}
|
|
|
|
bdebug("silent double reserve of PFN %lx\n",
|
2008-07-24 00:28:09 -04:00
|
|
|
idx + bdata->node_min_pfn);
|
2008-07-24 00:28:05 -04:00
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-07-24 00:28:06 -04:00
|
|
|
static int __init mark_bootmem_node(bootmem_data_t *bdata,
|
|
|
|
unsigned long start, unsigned long end,
|
|
|
|
int reserve, int flags)
|
2008-07-24 00:28:00 -04:00
|
|
|
{
|
|
|
|
unsigned long sidx, eidx;
|
|
|
|
|
2008-07-24 00:28:06 -04:00
|
|
|
bdebug("nid=%td start=%lx end=%lx reserve=%d flags=%x\n",
|
|
|
|
bdata - bootmem_node_data, start, end, reserve, flags);
|
2008-07-24 00:28:00 -04:00
|
|
|
|
2008-07-24 00:28:09 -04:00
|
|
|
BUG_ON(start < bdata->node_min_pfn);
|
2008-07-24 00:28:06 -04:00
|
|
|
BUG_ON(end > bdata->node_low_pfn);
|
2008-07-24 00:28:00 -04:00
|
|
|
|
2008-07-24 00:28:09 -04:00
|
|
|
sidx = start - bdata->node_min_pfn;
|
|
|
|
eidx = end - bdata->node_min_pfn;
|
2008-07-24 00:28:00 -04:00
|
|
|
|
2008-07-24 00:28:06 -04:00
|
|
|
if (reserve)
|
|
|
|
return __reserve(bdata, sidx, eidx, flags);
|
2008-07-24 00:28:00 -04:00
|
|
|
else
|
2008-07-24 00:28:06 -04:00
|
|
|
__free(bdata, sidx, eidx);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int __init mark_bootmem(unsigned long start, unsigned long end,
|
|
|
|
int reserve, int flags)
|
|
|
|
{
|
|
|
|
unsigned long pos;
|
|
|
|
bootmem_data_t *bdata;
|
|
|
|
|
|
|
|
pos = start;
|
|
|
|
list_for_each_entry(bdata, &bdata_list, list) {
|
|
|
|
int err;
|
|
|
|
unsigned long max;
|
|
|
|
|
2008-07-24 00:28:09 -04:00
|
|
|
if (pos < bdata->node_min_pfn ||
|
|
|
|
pos >= bdata->node_low_pfn) {
|
2008-07-24 00:28:06 -04:00
|
|
|
BUG_ON(pos != start);
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
max = min(bdata->node_low_pfn, end);
|
2008-07-24 00:28:00 -04:00
|
|
|
|
2008-07-24 00:28:06 -04:00
|
|
|
err = mark_bootmem_node(bdata, pos, max, reserve, flags);
|
|
|
|
if (reserve && err) {
|
|
|
|
mark_bootmem(start, pos, 0, 0);
|
|
|
|
return err;
|
|
|
|
}
|
2008-07-24 00:28:00 -04:00
|
|
|
|
2008-07-24 00:28:06 -04:00
|
|
|
if (max == end)
|
|
|
|
return 0;
|
|
|
|
pos = bdata->node_low_pfn;
|
|
|
|
}
|
|
|
|
BUG();
|
2008-07-24 00:28:00 -04:00
|
|
|
}
|
|
|
|
|
2008-07-24 00:28:01 -04:00
|
|
|
/**
|
|
|
|
* free_bootmem_node - mark a page range as usable
|
|
|
|
* @pgdat: node the range resides on
|
|
|
|
* @physaddr: starting address of the range
|
|
|
|
* @size: size of the range in bytes
|
|
|
|
*
|
|
|
|
* Partial pages will be considered reserved and left as they are.
|
|
|
|
*
|
2008-07-24 00:28:06 -04:00
|
|
|
* The range must reside completely on the specified node.
|
2008-07-24 00:28:01 -04:00
|
|
|
*/
|
2008-07-24 00:28:00 -04:00
|
|
|
void __init free_bootmem_node(pg_data_t *pgdat, unsigned long physaddr,
|
|
|
|
unsigned long size)
|
|
|
|
{
|
2008-07-24 00:28:06 -04:00
|
|
|
unsigned long start, end;
|
|
|
|
|
|
|
|
start = PFN_UP(physaddr);
|
|
|
|
end = PFN_DOWN(physaddr + size);
|
|
|
|
|
|
|
|
mark_bootmem_node(pgdat->bdata, start, end, 0, 0);
|
2008-07-24 00:28:00 -04:00
|
|
|
}
|
|
|
|
|
2008-07-24 00:28:01 -04:00
|
|
|
/**
|
|
|
|
* free_bootmem - mark a page range as usable
|
|
|
|
* @addr: starting address of the range
|
|
|
|
* @size: size of the range in bytes
|
|
|
|
*
|
|
|
|
* Partial pages will be considered reserved and left as they are.
|
|
|
|
*
|
2008-07-24 00:28:06 -04:00
|
|
|
* The range must be contiguous but may span node boundaries.
|
2008-07-24 00:28:01 -04:00
|
|
|
*/
|
2008-07-24 00:28:00 -04:00
|
|
|
void __init free_bootmem(unsigned long addr, unsigned long size)
|
|
|
|
{
|
2008-07-24 00:28:06 -04:00
|
|
|
unsigned long start, end;
|
2008-03-18 15:49:12 -04:00
|
|
|
|
2008-07-24 00:28:06 -04:00
|
|
|
start = PFN_UP(addr);
|
|
|
|
end = PFN_DOWN(addr + size);
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2008-07-24 00:28:06 -04:00
|
|
|
mark_bootmem(start, end, 0, 0);
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
2008-07-24 00:28:01 -04:00
|
|
|
/**
|
|
|
|
* reserve_bootmem_node - mark a page range as reserved
|
|
|
|
* @pgdat: node the range resides on
|
|
|
|
* @physaddr: starting address of the range
|
|
|
|
* @size: size of the range in bytes
|
|
|
|
* @flags: reservation flags (see linux/bootmem.h)
|
|
|
|
*
|
|
|
|
* Partial pages will be reserved.
|
|
|
|
*
|
2008-07-24 00:28:06 -04:00
|
|
|
* The range must reside completely on the specified node.
|
2008-07-24 00:28:01 -04:00
|
|
|
*/
|
2008-07-24 00:28:00 -04:00
|
|
|
int __init reserve_bootmem_node(pg_data_t *pgdat, unsigned long physaddr,
|
|
|
|
unsigned long size, int flags)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
2008-07-24 00:28:06 -04:00
|
|
|
unsigned long start, end;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2008-07-24 00:28:06 -04:00
|
|
|
start = PFN_DOWN(physaddr);
|
|
|
|
end = PFN_UP(physaddr + size);
|
|
|
|
|
|
|
|
return mark_bootmem_node(pgdat->bdata, start, end, 1, flags);
|
2008-07-24 00:28:00 -04:00
|
|
|
}
|
2008-03-24 15:29:45 -04:00
|
|
|
|
2008-07-24 00:28:00 -04:00
|
|
|
#ifndef CONFIG_HAVE_ARCH_BOOTMEM_NODE
|
2008-07-24 00:28:01 -04:00
|
|
|
/**
|
|
|
|
* reserve_bootmem - mark a page range as usable
|
|
|
|
* @addr: starting address of the range
|
|
|
|
* @size: size of the range in bytes
|
|
|
|
* @flags: reservation flags (see linux/bootmem.h)
|
|
|
|
*
|
|
|
|
* Partial pages will be reserved.
|
|
|
|
*
|
2008-07-24 00:28:06 -04:00
|
|
|
* The range must be contiguous but may span node boundaries.
|
2008-07-24 00:28:01 -04:00
|
|
|
*/
|
2008-07-24 00:28:00 -04:00
|
|
|
int __init reserve_bootmem(unsigned long addr, unsigned long size,
|
|
|
|
int flags)
|
|
|
|
{
|
2008-07-24 00:28:06 -04:00
|
|
|
unsigned long start, end;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2008-07-24 00:28:06 -04:00
|
|
|
start = PFN_DOWN(addr);
|
|
|
|
end = PFN_UP(addr + size);
|
2008-07-24 00:28:00 -04:00
|
|
|
|
2008-07-24 00:28:06 -04:00
|
|
|
return mark_bootmem(start, end, 1, flags);
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
2008-07-24 00:28:00 -04:00
|
|
|
#endif /* !CONFIG_HAVE_ARCH_BOOTMEM_NODE */
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2008-08-20 17:09:15 -04:00
|
|
|
static unsigned long align_idx(struct bootmem_data *bdata, unsigned long idx,
|
|
|
|
unsigned long step)
|
|
|
|
{
|
|
|
|
unsigned long base = bdata->node_min_pfn;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Align the index with respect to the node start so that the
|
|
|
|
* combination of both satisfies the requested alignment.
|
|
|
|
*/
|
|
|
|
|
|
|
|
return ALIGN(base + idx, step) - base;
|
|
|
|
}
|
|
|
|
|
|
|
|
static unsigned long align_off(struct bootmem_data *bdata, unsigned long off,
|
|
|
|
unsigned long align)
|
|
|
|
{
|
|
|
|
unsigned long base = PFN_PHYS(bdata->node_min_pfn);
|
|
|
|
|
|
|
|
/* Same as align_idx for byte offsets */
|
|
|
|
|
|
|
|
return ALIGN(base + off, align) - base;
|
|
|
|
}
|
|
|
|
|
2008-07-24 00:28:05 -04:00
|
|
|
static void * __init alloc_bootmem_core(struct bootmem_data *bdata,
|
|
|
|
unsigned long size, unsigned long align,
|
|
|
|
unsigned long goal, unsigned long limit)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
2008-07-24 00:28:07 -04:00
|
|
|
unsigned long fallback = 0;
|
2008-07-24 00:28:05 -04:00
|
|
|
unsigned long min, max, start, sidx, midx, step;
|
|
|
|
|
2009-01-06 17:40:32 -05:00
|
|
|
bdebug("nid=%td size=%lx [%lu pages] align=%lx goal=%lx limit=%lx\n",
|
|
|
|
bdata - bootmem_node_data, size, PAGE_ALIGN(size) >> PAGE_SHIFT,
|
|
|
|
align, goal, limit);
|
|
|
|
|
2008-07-24 00:28:05 -04:00
|
|
|
BUG_ON(!size);
|
|
|
|
BUG_ON(align & (align - 1));
|
|
|
|
BUG_ON(limit && goal + size > limit);
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2006-12-06 23:32:41 -05:00
|
|
|
if (!bdata->node_bootmem_map)
|
|
|
|
return NULL;
|
|
|
|
|
2008-07-24 00:28:09 -04:00
|
|
|
min = bdata->node_min_pfn;
|
2008-07-24 00:28:05 -04:00
|
|
|
max = bdata->node_low_pfn;
|
2008-03-18 15:44:48 -04:00
|
|
|
|
2008-07-24 00:28:05 -04:00
|
|
|
goal >>= PAGE_SHIFT;
|
|
|
|
limit >>= PAGE_SHIFT;
|
|
|
|
|
|
|
|
if (limit && max > limit)
|
|
|
|
max = limit;
|
|
|
|
if (max <= min)
|
2008-03-18 15:44:48 -04:00
|
|
|
return NULL;
|
|
|
|
|
2008-07-24 00:28:05 -04:00
|
|
|
step = max(align >> PAGE_SHIFT, 1UL);
|
2005-10-19 18:52:18 -04:00
|
|
|
|
2008-07-24 00:28:05 -04:00
|
|
|
if (goal && min < goal && goal < max)
|
|
|
|
start = ALIGN(goal, step);
|
|
|
|
else
|
|
|
|
start = ALIGN(min, step);
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2008-08-20 17:09:15 -04:00
|
|
|
sidx = start - bdata->node_min_pfn;
|
2008-07-24 00:28:09 -04:00
|
|
|
midx = max - bdata->node_min_pfn;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2008-07-24 00:28:05 -04:00
|
|
|
if (bdata->hint_idx > sidx) {
|
2008-07-24 00:28:07 -04:00
|
|
|
/*
|
|
|
|
* Handle the valid case of sidx being zero and still
|
|
|
|
* catch the fallback below.
|
|
|
|
*/
|
|
|
|
fallback = sidx + 1;
|
2008-08-20 17:09:15 -04:00
|
|
|
sidx = align_idx(bdata, bdata->hint_idx, step);
|
2008-07-24 00:28:05 -04:00
|
|
|
}
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2008-07-24 00:28:05 -04:00
|
|
|
while (1) {
|
|
|
|
int merge;
|
|
|
|
void *region;
|
|
|
|
unsigned long eidx, i, start_off, end_off;
|
|
|
|
find_block:
|
|
|
|
sidx = find_next_zero_bit(bdata->node_bootmem_map, midx, sidx);
|
2008-08-20 17:09:15 -04:00
|
|
|
sidx = align_idx(bdata, sidx, step);
|
2008-07-24 00:28:05 -04:00
|
|
|
eidx = sidx + PFN_UP(size);
|
2008-03-11 02:23:42 -04:00
|
|
|
|
2008-07-24 00:28:05 -04:00
|
|
|
if (sidx >= midx || eidx > midx)
|
2005-12-12 03:37:39 -05:00
|
|
|
break;
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2008-07-24 00:28:05 -04:00
|
|
|
for (i = sidx; i < eidx; i++)
|
|
|
|
if (test_bit(i, bdata->node_bootmem_map)) {
|
2008-08-20 17:09:15 -04:00
|
|
|
sidx = align_idx(bdata, i, step);
|
2008-07-24 00:28:05 -04:00
|
|
|
if (sidx == i)
|
|
|
|
sidx += step;
|
|
|
|
goto find_block;
|
|
|
|
}
|
2005-04-16 18:20:36 -04:00
|
|
|
|
bootmem allocator: alloc_bootmem_core(): page-align the end offset
This is the minimal sequence that jams the allocator:
void *p, *q, *r;
p = alloc_bootmem(PAGE_SIZE);
q = alloc_bootmem(64);
free_bootmem(p, PAGE_SIZE);
p = alloc_bootmem(PAGE_SIZE);
r = alloc_bootmem(64);
after this sequence (assuming that the allocator was empty or page-aligned
before), pointer "q" will be equal to pointer "r".
What's hapenning inside the allocator:
p = alloc_bootmem(PAGE_SIZE);
in allocator: last_end_off == PAGE_SIZE, bitmap contains bits 10000...
q = alloc_bootmem(64);
in allocator: last_end_off == PAGE_SIZE + 64, bitmap contains 11000...
free_bootmem(p, PAGE_SIZE);
in allocator: last_end_off == PAGE_SIZE + 64, bitmap contains 01000...
p = alloc_bootmem(PAGE_SIZE);
in allocator: last_end_off == PAGE_SIZE, bitmap contains 11000...
r = alloc_bootmem(64);
and now:
it finds bit "2", as a place where to allocate (sidx)
it hits the condition
if (bdata->last_end_off && PFN_DOWN(bdata->last_end_off) + 1 == sidx))
start_off = ALIGN(bdata->last_end_off, align);
-you can see that the condition is true, so it assigns start_off =
ALIGN(bdata->last_end_off, align); (that is PAGE_SIZE) and allocates
over already allocated block.
With the patch it tries to continue at the end of previous allocation only
if the previous allocation ended in the middle of the page.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Acked-by: Johannes Weiner <hannes@saeurebad.de>
Cc: David Miller <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-08-15 03:40:17 -04:00
|
|
|
if (bdata->last_end_off & (PAGE_SIZE - 1) &&
|
2008-07-24 00:28:05 -04:00
|
|
|
PFN_DOWN(bdata->last_end_off) + 1 == sidx)
|
2008-08-20 17:09:15 -04:00
|
|
|
start_off = align_off(bdata, bdata->last_end_off, align);
|
2008-07-24 00:28:05 -04:00
|
|
|
else
|
|
|
|
start_off = PFN_PHYS(sidx);
|
|
|
|
|
|
|
|
merge = PFN_DOWN(start_off) < sidx;
|
|
|
|
end_off = start_off + size;
|
|
|
|
|
|
|
|
bdata->last_end_off = end_off;
|
|
|
|
bdata->hint_idx = PFN_UP(end_off);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Reserve the area now:
|
|
|
|
*/
|
2008-07-24 00:28:05 -04:00
|
|
|
if (__reserve(bdata, PFN_DOWN(start_off) + merge,
|
|
|
|
PFN_UP(end_off), BOOTMEM_EXCLUSIVE))
|
|
|
|
BUG();
|
2008-07-24 00:28:05 -04:00
|
|
|
|
2008-07-24 00:28:09 -04:00
|
|
|
region = phys_to_virt(PFN_PHYS(bdata->node_min_pfn) +
|
|
|
|
start_off);
|
2008-07-24 00:28:05 -04:00
|
|
|
memset(region, 0, size);
|
|
|
|
return region;
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
2008-07-24 00:28:07 -04:00
|
|
|
if (fallback) {
|
2008-08-20 17:09:15 -04:00
|
|
|
sidx = align_idx(bdata, fallback - 1, step);
|
2008-07-24 00:28:07 -04:00
|
|
|
fallback = 0;
|
|
|
|
goto find_block;
|
|
|
|
}
|
|
|
|
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void * __init ___alloc_bootmem_nopanic(unsigned long size,
|
|
|
|
unsigned long align,
|
|
|
|
unsigned long goal,
|
|
|
|
unsigned long limit)
|
|
|
|
{
|
|
|
|
bootmem_data_t *bdata;
|
|
|
|
|
|
|
|
restart:
|
|
|
|
list_for_each_entry(bdata, &bdata_list, list) {
|
|
|
|
void *region;
|
|
|
|
|
|
|
|
if (goal && bdata->node_low_pfn <= PFN_DOWN(goal))
|
|
|
|
continue;
|
2008-07-24 00:28:09 -04:00
|
|
|
if (limit && bdata->node_min_pfn >= PFN_DOWN(limit))
|
2008-07-24 00:28:07 -04:00
|
|
|
break;
|
|
|
|
|
|
|
|
region = alloc_bootmem_core(bdata, size, align, goal, limit);
|
|
|
|
if (region)
|
|
|
|
return region;
|
|
|
|
}
|
|
|
|
|
2008-07-24 00:28:05 -04:00
|
|
|
if (goal) {
|
|
|
|
goal = 0;
|
2008-07-24 00:28:07 -04:00
|
|
|
goto restart;
|
2008-07-24 00:28:05 -04:00
|
|
|
}
|
2008-07-24 00:28:02 -04:00
|
|
|
|
2008-07-24 00:28:05 -04:00
|
|
|
return NULL;
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
2008-07-24 00:28:01 -04:00
|
|
|
/**
|
|
|
|
* __alloc_bootmem_nopanic - allocate boot memory without panicking
|
|
|
|
* @size: size of the request in bytes
|
|
|
|
* @align: alignment of the region
|
|
|
|
* @goal: preferred starting address of the region
|
|
|
|
*
|
|
|
|
* The goal is dropped if it can not be satisfied and the allocation will
|
|
|
|
* fall back to memory below @goal.
|
|
|
|
*
|
|
|
|
* Allocation may happen on any node in the system.
|
|
|
|
*
|
|
|
|
* Returns NULL on failure.
|
|
|
|
*/
|
2006-09-26 02:31:05 -04:00
|
|
|
void * __init __alloc_bootmem_nopanic(unsigned long size, unsigned long align,
|
2008-07-24 00:28:07 -04:00
|
|
|
unsigned long goal)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
2008-07-24 00:28:07 -04:00
|
|
|
return ___alloc_bootmem_nopanic(size, align, goal, 0);
|
|
|
|
}
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2008-07-24 00:28:07 -04:00
|
|
|
static void * __init ___alloc_bootmem(unsigned long size, unsigned long align,
|
|
|
|
unsigned long goal, unsigned long limit)
|
|
|
|
{
|
|
|
|
void *mem = ___alloc_bootmem_nopanic(size, align, goal, limit);
|
|
|
|
|
|
|
|
if (mem)
|
|
|
|
return mem;
|
|
|
|
/*
|
|
|
|
* Whoops, we cannot satisfy the allocation request.
|
|
|
|
*/
|
|
|
|
printk(KERN_ALERT "bootmem alloc of %lu bytes failed!\n", size);
|
|
|
|
panic("Out of memory");
|
2006-04-07 13:49:21 -04:00
|
|
|
return NULL;
|
|
|
|
}
|
2005-04-16 18:20:36 -04:00
|
|
|
|
2008-07-24 00:28:01 -04:00
|
|
|
/**
|
|
|
|
* __alloc_bootmem - allocate boot memory
|
|
|
|
* @size: size of the request in bytes
|
|
|
|
* @align: alignment of the region
|
|
|
|
* @goal: preferred starting address of the region
|
|
|
|
*
|
|
|
|
* The goal is dropped if it can not be satisfied and the allocation will
|
|
|
|
* fall back to memory below @goal.
|
|
|
|
*
|
|
|
|
* Allocation may happen on any node in the system.
|
|
|
|
*
|
|
|
|
* The function panics if the request can not be satisfied.
|
|
|
|
*/
|
2006-09-26 02:31:05 -04:00
|
|
|
void * __init __alloc_bootmem(unsigned long size, unsigned long align,
|
|
|
|
unsigned long goal)
|
2006-04-07 13:49:21 -04:00
|
|
|
{
|
2008-07-24 00:28:07 -04:00
|
|
|
return ___alloc_bootmem(size, align, goal, 0);
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
2008-07-24 00:28:08 -04:00
|
|
|
static void * __init ___alloc_bootmem_node(bootmem_data_t *bdata,
|
|
|
|
unsigned long size, unsigned long align,
|
|
|
|
unsigned long goal, unsigned long limit)
|
|
|
|
{
|
|
|
|
void *ptr;
|
|
|
|
|
|
|
|
ptr = alloc_bootmem_core(bdata, size, align, goal, limit);
|
|
|
|
if (ptr)
|
|
|
|
return ptr;
|
|
|
|
|
|
|
|
return ___alloc_bootmem(size, align, goal, limit);
|
|
|
|
}
|
|
|
|
|
2008-07-24 00:28:01 -04:00
|
|
|
/**
|
|
|
|
* __alloc_bootmem_node - allocate boot memory from a specific node
|
|
|
|
* @pgdat: node to allocate from
|
|
|
|
* @size: size of the request in bytes
|
|
|
|
* @align: alignment of the region
|
|
|
|
* @goal: preferred starting address of the region
|
|
|
|
*
|
|
|
|
* The goal is dropped if it can not be satisfied and the allocation will
|
|
|
|
* fall back to memory below @goal.
|
|
|
|
*
|
|
|
|
* Allocation may fall back to any node in the system if the specified node
|
|
|
|
* can not hold the requested memory.
|
|
|
|
*
|
|
|
|
* The function panics if the request can not be satisfied.
|
|
|
|
*/
|
2006-09-26 02:31:05 -04:00
|
|
|
void * __init __alloc_bootmem_node(pg_data_t *pgdat, unsigned long size,
|
|
|
|
unsigned long align, unsigned long goal)
|
2005-04-16 18:20:36 -04:00
|
|
|
{
|
2008-07-24 00:28:08 -04:00
|
|
|
return ___alloc_bootmem_node(pgdat->bdata, size, align, goal, 0);
|
2005-04-16 18:20:36 -04:00
|
|
|
}
|
|
|
|
|
2008-04-28 05:13:32 -04:00
|
|
|
#ifdef CONFIG_SPARSEMEM
|
2008-07-24 00:28:01 -04:00
|
|
|
/**
|
|
|
|
* alloc_bootmem_section - allocate boot memory from a specific section
|
|
|
|
* @size: size of the request in bytes
|
|
|
|
* @section_nr: sparse map section to allocate from
|
|
|
|
*
|
|
|
|
* Return NULL on failure.
|
|
|
|
*/
|
2008-04-28 05:13:32 -04:00
|
|
|
void * __init alloc_bootmem_section(unsigned long size,
|
|
|
|
unsigned long section_nr)
|
|
|
|
{
|
2008-07-24 00:28:09 -04:00
|
|
|
bootmem_data_t *bdata;
|
|
|
|
unsigned long pfn, goal, limit;
|
2008-04-28 05:13:32 -04:00
|
|
|
|
|
|
|
pfn = section_nr_to_pfn(section_nr);
|
2008-07-24 00:28:09 -04:00
|
|
|
goal = pfn << PAGE_SHIFT;
|
|
|
|
limit = section_nr_to_pfn(section_nr + 1) << PAGE_SHIFT;
|
|
|
|
bdata = &bootmem_node_data[early_pfn_to_nid(pfn)];
|
2008-04-28 05:13:32 -04:00
|
|
|
|
2008-07-24 00:28:09 -04:00
|
|
|
return alloc_bootmem_core(bdata, size, SMP_CACHE_BYTES, goal, limit);
|
2008-04-28 05:13:32 -04:00
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2008-07-24 00:27:45 -04:00
|
|
|
void * __init __alloc_bootmem_node_nopanic(pg_data_t *pgdat, unsigned long size,
|
|
|
|
unsigned long align, unsigned long goal)
|
|
|
|
{
|
|
|
|
void *ptr;
|
|
|
|
|
|
|
|
ptr = alloc_bootmem_core(pgdat->bdata, size, align, goal, 0);
|
|
|
|
if (ptr)
|
|
|
|
return ptr;
|
|
|
|
|
|
|
|
return __alloc_bootmem_nopanic(size, align, goal);
|
|
|
|
}
|
|
|
|
|
2006-09-26 02:31:33 -04:00
|
|
|
#ifndef ARCH_LOW_ADDRESS_LIMIT
|
|
|
|
#define ARCH_LOW_ADDRESS_LIMIT 0xffffffffUL
|
|
|
|
#endif
|
2006-01-06 03:11:01 -05:00
|
|
|
|
2008-07-24 00:28:01 -04:00
|
|
|
/**
|
|
|
|
* __alloc_bootmem_low - allocate low boot memory
|
|
|
|
* @size: size of the request in bytes
|
|
|
|
* @align: alignment of the region
|
|
|
|
* @goal: preferred starting address of the region
|
|
|
|
*
|
|
|
|
* The goal is dropped if it can not be satisfied and the allocation will
|
|
|
|
* fall back to memory below @goal.
|
|
|
|
*
|
|
|
|
* Allocation may happen on any node in the system.
|
|
|
|
*
|
|
|
|
* The function panics if the request can not be satisfied.
|
|
|
|
*/
|
2006-09-26 02:31:05 -04:00
|
|
|
void * __init __alloc_bootmem_low(unsigned long size, unsigned long align,
|
|
|
|
unsigned long goal)
|
2006-01-06 03:11:01 -05:00
|
|
|
{
|
2008-07-24 00:28:07 -04:00
|
|
|
return ___alloc_bootmem(size, align, goal, ARCH_LOW_ADDRESS_LIMIT);
|
2006-01-06 03:11:01 -05:00
|
|
|
}
|
|
|
|
|
2008-07-24 00:28:01 -04:00
|
|
|
/**
|
|
|
|
* __alloc_bootmem_low_node - allocate low boot memory from a specific node
|
|
|
|
* @pgdat: node to allocate from
|
|
|
|
* @size: size of the request in bytes
|
|
|
|
* @align: alignment of the region
|
|
|
|
* @goal: preferred starting address of the region
|
|
|
|
*
|
|
|
|
* The goal is dropped if it can not be satisfied and the allocation will
|
|
|
|
* fall back to memory below @goal.
|
|
|
|
*
|
|
|
|
* Allocation may fall back to any node in the system if the specified node
|
|
|
|
* can not hold the requested memory.
|
|
|
|
*
|
|
|
|
* The function panics if the request can not be satisfied.
|
|
|
|
*/
|
2006-01-06 03:11:01 -05:00
|
|
|
void * __init __alloc_bootmem_low_node(pg_data_t *pgdat, unsigned long size,
|
|
|
|
unsigned long align, unsigned long goal)
|
|
|
|
{
|
2008-07-24 00:28:08 -04:00
|
|
|
return ___alloc_bootmem_node(pgdat->bdata, size, align,
|
|
|
|
goal, ARCH_LOW_ADDRESS_LIMIT);
|
2006-01-06 03:11:01 -05:00
|
|
|
}
|