linux/mm/hugetlb.c

799 lines
18 KiB
C
Raw Normal View History

/*
* Generic hugetlb support.
* (C) William Irwin, April 2004
*/
#include <linux/gfp.h>
#include <linux/list.h>
#include <linux/init.h>
#include <linux/module.h>
#include <linux/mm.h>
#include <linux/sysctl.h>
#include <linux/highmem.h>
#include <linux/nodemask.h>
#include <linux/pagemap.h>
#include <linux/mempolicy.h>
#include <linux/cpuset.h>
[PATCH] hugepage: serialize hugepage allocation and instantiation Currently, no lock or mutex is held between allocating a hugepage and inserting it into the pagetables / page cache. When we do go to insert the page into pagetables or page cache, we recheck and may free the newly allocated hugepage. However, since the number of hugepages in the system is strictly limited, and it's usualy to want to use all of them, this can still lead to spurious allocation failures. For example, suppose two processes are both mapping (MAP_SHARED) the same hugepage file, large enough to consume the entire available hugepage pool. If they race instantiating the last page in the mapping, they will both attempt to allocate the last available hugepage. One will fail, of course, returning OOM from the fault and thus causing the process to be killed, despite the fact that the entire mapping can, in fact, be instantiated. The patch fixes this race by the simple method of adding a (sleeping) mutex to serialize the hugepage fault path between allocation and insertion into pagetables and/or page cache. It would be possible to avoid the serialization by catching the allocation failures, waiting on some condition, then rechecking to see if someone else has instantiated the page for us. Given the likely frequency of hugepage instantiations, it seems very doubtful it's worth the extra complexity. This patch causes no regression on the libhugetlbfs testsuite, and one test, which can trigger this race now passes where it previously failed. Actually, the test still sometimes fails, though less often and only as a shmat() failure, rather processes getting OOM killed by the VM. The dodgy heuristic tests in fs/hugetlbfs/inode.c for whether there's enough hugepage space aren't protected by the new mutex, and would be ugly to do so, so there's still a race there. Another patch to replace those tests with something saner for this reason as well as others coming... Signed-off-by: David Gibson <dwg@au1.ibm.com> Cc: William Lee Irwin III <wli@holomorphy.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-22 08:08:53 +00:00
#include <linux/mutex.h>
#include <asm/page.h>
#include <asm/pgtable.h>
#include <linux/hugetlb.h>
#include "internal.h"
const unsigned long hugetlb_zero = 0, hugetlb_infinity = ~0UL;
static unsigned long nr_huge_pages, free_huge_pages, resv_huge_pages;
unsigned long max_huge_pages;
static struct list_head hugepage_freelists[MAX_NUMNODES];
static unsigned int nr_huge_pages_node[MAX_NUMNODES];
static unsigned int free_huge_pages_node[MAX_NUMNODES];
[PATCH] hugepage: serialize hugepage allocation and instantiation Currently, no lock or mutex is held between allocating a hugepage and inserting it into the pagetables / page cache. When we do go to insert the page into pagetables or page cache, we recheck and may free the newly allocated hugepage. However, since the number of hugepages in the system is strictly limited, and it's usualy to want to use all of them, this can still lead to spurious allocation failures. For example, suppose two processes are both mapping (MAP_SHARED) the same hugepage file, large enough to consume the entire available hugepage pool. If they race instantiating the last page in the mapping, they will both attempt to allocate the last available hugepage. One will fail, of course, returning OOM from the fault and thus causing the process to be killed, despite the fact that the entire mapping can, in fact, be instantiated. The patch fixes this race by the simple method of adding a (sleeping) mutex to serialize the hugepage fault path between allocation and insertion into pagetables and/or page cache. It would be possible to avoid the serialization by catching the allocation failures, waiting on some condition, then rechecking to see if someone else has instantiated the page for us. Given the likely frequency of hugepage instantiations, it seems very doubtful it's worth the extra complexity. This patch causes no regression on the libhugetlbfs testsuite, and one test, which can trigger this race now passes where it previously failed. Actually, the test still sometimes fails, though less often and only as a shmat() failure, rather processes getting OOM killed by the VM. The dodgy heuristic tests in fs/hugetlbfs/inode.c for whether there's enough hugepage space aren't protected by the new mutex, and would be ugly to do so, so there's still a race there. Another patch to replace those tests with something saner for this reason as well as others coming... Signed-off-by: David Gibson <dwg@au1.ibm.com> Cc: William Lee Irwin III <wli@holomorphy.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-22 08:08:53 +00:00
/*
* Protects updates to hugepage_freelists, nr_huge_pages, and free_huge_pages
*/
static DEFINE_SPINLOCK(hugetlb_lock);
static void clear_huge_page(struct page *page, unsigned long addr)
{
int i;
might_sleep();
for (i = 0; i < (HPAGE_SIZE/PAGE_SIZE); i++) {
cond_resched();
clear_user_highpage(page + i, addr);
}
}
static void copy_huge_page(struct page *dst, struct page *src,
unsigned long addr)
{
int i;
might_sleep();
for (i = 0; i < HPAGE_SIZE/PAGE_SIZE; i++) {
cond_resched();
copy_user_highpage(dst + i, src + i, addr + i*PAGE_SIZE);
}
}
static void enqueue_huge_page(struct page *page)
{
int nid = page_to_nid(page);
list_add(&page->lru, &hugepage_freelists[nid]);
free_huge_pages++;
free_huge_pages_node[nid]++;
}
static struct page *dequeue_huge_page(struct vm_area_struct *vma,
unsigned long address)
{
int nid = numa_node_id();
struct page *page = NULL;
struct zonelist *zonelist = huge_zonelist(vma, address);
struct zone **z;
for (z = zonelist->zones; *z; z++) {
nid = (*z)->zone_pgdat->node_id;
if (cpuset_zone_allowed(*z, GFP_HIGHUSER) &&
!list_empty(&hugepage_freelists[nid]))
break;
}
if (*z) {
page = list_entry(hugepage_freelists[nid].next,
struct page, lru);
list_del(&page->lru);
free_huge_pages--;
free_huge_pages_node[nid]--;
}
return page;
}
static void free_huge_page(struct page *page)
{
BUG_ON(page_count(page));
INIT_LIST_HEAD(&page->lru);
spin_lock(&hugetlb_lock);
enqueue_huge_page(page);
spin_unlock(&hugetlb_lock);
}
static int alloc_fresh_huge_page(void)
{
static int nid = 0;
struct page *page;
page = alloc_pages_node(nid, GFP_HIGHUSER|__GFP_COMP|__GFP_NOWARN,
HUGETLB_PAGE_ORDER);
nid = next_node(nid, node_online_map);
if (nid == MAX_NUMNODES)
nid = first_node(node_online_map);
if (page) {
page[1].lru.next = (void *)free_huge_page; /* dtor */
spin_lock(&hugetlb_lock);
nr_huge_pages++;
nr_huge_pages_node[page_to_nid(page)]++;
spin_unlock(&hugetlb_lock);
put_page(page); /* free it into the hugepage allocator */
return 1;
}
return 0;
}
static struct page *alloc_huge_page(struct vm_area_struct *vma,
unsigned long addr)
{
struct page *page;
spin_lock(&hugetlb_lock);
if (vma->vm_flags & VM_MAYSHARE)
resv_huge_pages--;
else if (free_huge_pages <= resv_huge_pages)
goto fail;
[PATCH] hugepage: Strict page reservation for hugepage inodes These days, hugepages are demand-allocated at first fault time. There's a somewhat dubious (and racy) heuristic when making a new mmap() to check if there are enough available hugepages to fully satisfy that mapping. A particularly obvious case where the heuristic breaks down is where a process maps its hugepages not as a single chunk, but as a bunch of individually mmap()ed (or shmat()ed) blocks without touching and instantiating the pages in between allocations. In this case the size of each block is compared against the total number of available hugepages. It's thus easy for the process to become overcommitted, because each block mapping will succeed, although the total number of hugepages required by all blocks exceeds the number available. In particular, this defeats such a program which will detect a mapping failure and adjust its hugepage usage downward accordingly. The patch below addresses this problem, by strictly reserving a number of physical hugepages for hugepage inodes which have been mapped, but not instatiated. MAP_SHARED mappings are thus "safe" - they will fail on mmap(), not later with an OOM SIGKILL. MAP_PRIVATE mappings can still trigger an OOM. (Actually SHARED mappings can technically still OOM, but only if the sysadmin explicitly reduces the hugepage pool between mapping and instantiation) This patch appears to address the problem at hand - it allows DB2 to start correctly, for instance, which previously suffered the failure described above. This patch causes no regressions on the libhugetblfs testsuite, and makes a test (designed to catch this problem) pass which previously failed (ppc64, POWER5). Signed-off-by: David Gibson <dwg@au1.ibm.com> Cc: William Lee Irwin III <wli@holomorphy.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-22 08:08:55 +00:00
page = dequeue_huge_page(vma, addr);
if (!page)
goto fail;
spin_unlock(&hugetlb_lock);
set_page_refcounted(page);
return page;
[PATCH] hugepage: Strict page reservation for hugepage inodes These days, hugepages are demand-allocated at first fault time. There's a somewhat dubious (and racy) heuristic when making a new mmap() to check if there are enough available hugepages to fully satisfy that mapping. A particularly obvious case where the heuristic breaks down is where a process maps its hugepages not as a single chunk, but as a bunch of individually mmap()ed (or shmat()ed) blocks without touching and instantiating the pages in between allocations. In this case the size of each block is compared against the total number of available hugepages. It's thus easy for the process to become overcommitted, because each block mapping will succeed, although the total number of hugepages required by all blocks exceeds the number available. In particular, this defeats such a program which will detect a mapping failure and adjust its hugepage usage downward accordingly. The patch below addresses this problem, by strictly reserving a number of physical hugepages for hugepage inodes which have been mapped, but not instatiated. MAP_SHARED mappings are thus "safe" - they will fail on mmap(), not later with an OOM SIGKILL. MAP_PRIVATE mappings can still trigger an OOM. (Actually SHARED mappings can technically still OOM, but only if the sysadmin explicitly reduces the hugepage pool between mapping and instantiation) This patch appears to address the problem at hand - it allows DB2 to start correctly, for instance, which previously suffered the failure described above. This patch causes no regressions on the libhugetblfs testsuite, and makes a test (designed to catch this problem) pass which previously failed (ppc64, POWER5). Signed-off-by: David Gibson <dwg@au1.ibm.com> Cc: William Lee Irwin III <wli@holomorphy.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-22 08:08:55 +00:00
fail:
[PATCH] hugepage: Strict page reservation for hugepage inodes These days, hugepages are demand-allocated at first fault time. There's a somewhat dubious (and racy) heuristic when making a new mmap() to check if there are enough available hugepages to fully satisfy that mapping. A particularly obvious case where the heuristic breaks down is where a process maps its hugepages not as a single chunk, but as a bunch of individually mmap()ed (or shmat()ed) blocks without touching and instantiating the pages in between allocations. In this case the size of each block is compared against the total number of available hugepages. It's thus easy for the process to become overcommitted, because each block mapping will succeed, although the total number of hugepages required by all blocks exceeds the number available. In particular, this defeats such a program which will detect a mapping failure and adjust its hugepage usage downward accordingly. The patch below addresses this problem, by strictly reserving a number of physical hugepages for hugepage inodes which have been mapped, but not instatiated. MAP_SHARED mappings are thus "safe" - they will fail on mmap(), not later with an OOM SIGKILL. MAP_PRIVATE mappings can still trigger an OOM. (Actually SHARED mappings can technically still OOM, but only if the sysadmin explicitly reduces the hugepage pool between mapping and instantiation) This patch appears to address the problem at hand - it allows DB2 to start correctly, for instance, which previously suffered the failure described above. This patch causes no regressions on the libhugetblfs testsuite, and makes a test (designed to catch this problem) pass which previously failed (ppc64, POWER5). Signed-off-by: David Gibson <dwg@au1.ibm.com> Cc: William Lee Irwin III <wli@holomorphy.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-22 08:08:55 +00:00
spin_unlock(&hugetlb_lock);
return NULL;
}
static int __init hugetlb_init(void)
{
unsigned long i;
if (HPAGE_SHIFT == 0)
return 0;
for (i = 0; i < MAX_NUMNODES; ++i)
INIT_LIST_HEAD(&hugepage_freelists[i]);
for (i = 0; i < max_huge_pages; ++i) {
if (!alloc_fresh_huge_page())
break;
}
max_huge_pages = free_huge_pages = nr_huge_pages = i;
printk("Total HugeTLB memory allocated, %ld\n", free_huge_pages);
return 0;
}
module_init(hugetlb_init);
static int __init hugetlb_setup(char *s)
{
if (sscanf(s, "%lu", &max_huge_pages) <= 0)
max_huge_pages = 0;
return 1;
}
__setup("hugepages=", hugetlb_setup);
#ifdef CONFIG_SYSCTL
static void update_and_free_page(struct page *page)
{
int i;
nr_huge_pages--;
nr_huge_pages_node[page_zone(page)->zone_pgdat->node_id]--;
for (i = 0; i < (HPAGE_SIZE / PAGE_SIZE); i++) {
page[i].flags &= ~(1 << PG_locked | 1 << PG_error | 1 << PG_referenced |
1 << PG_dirty | 1 << PG_active | 1 << PG_reserved |
1 << PG_private | 1<< PG_writeback);
}
page[1].lru.next = NULL;
set_page_refcounted(page);
__free_pages(page, HUGETLB_PAGE_ORDER);
}
#ifdef CONFIG_HIGHMEM
static void try_to_free_low(unsigned long count)
{
int i, nid;
for (i = 0; i < MAX_NUMNODES; ++i) {
struct page *page, *next;
list_for_each_entry_safe(page, next, &hugepage_freelists[i], lru) {
if (PageHighMem(page))
continue;
list_del(&page->lru);
update_and_free_page(page);
nid = page_zone(page)->zone_pgdat->node_id;
free_huge_pages--;
free_huge_pages_node[nid]--;
if (count >= nr_huge_pages)
return;
}
}
}
#else
static inline void try_to_free_low(unsigned long count)
{
}
#endif
static unsigned long set_max_huge_pages(unsigned long count)
{
while (count > nr_huge_pages) {
if (!alloc_fresh_huge_page())
return nr_huge_pages;
}
if (count >= nr_huge_pages)
return nr_huge_pages;
spin_lock(&hugetlb_lock);
count = max(count, resv_huge_pages);
try_to_free_low(count);
while (count < nr_huge_pages) {
struct page *page = dequeue_huge_page(NULL, 0);
if (!page)
break;
update_and_free_page(page);
}
spin_unlock(&hugetlb_lock);
return nr_huge_pages;
}
int hugetlb_sysctl_handler(struct ctl_table *table, int write,
struct file *file, void __user *buffer,
size_t *length, loff_t *ppos)
{
proc_doulongvec_minmax(table, write, file, buffer, length, ppos);
max_huge_pages = set_max_huge_pages(max_huge_pages);
return 0;
}
#endif /* CONFIG_SYSCTL */
int hugetlb_report_meminfo(char *buf)
{
return sprintf(buf,
"HugePages_Total: %5lu\n"
"HugePages_Free: %5lu\n"
"HugePages_Rsvd: %5lu\n"
"Hugepagesize: %5lu kB\n",
nr_huge_pages,
free_huge_pages,
resv_huge_pages,
HPAGE_SIZE/1024);
}
int hugetlb_report_node_meminfo(int nid, char *buf)
{
return sprintf(buf,
"Node %d HugePages_Total: %5u\n"
"Node %d HugePages_Free: %5u\n",
nid, nr_huge_pages_node[nid],
nid, free_huge_pages_node[nid]);
}
/* Return the number pages of memory we physically have, in PAGE_SIZE units. */
unsigned long hugetlb_total_pages(void)
{
return nr_huge_pages * (HPAGE_SIZE / PAGE_SIZE);
}
/*
* We cannot handle pagefaults against hugetlb pages at all. They cause
* handle_mm_fault() to try to instantiate regular-sized pages in the
* hugegpage VMA. do_page_fault() is supposed to trap this, so BUG is we get
* this far.
*/
static struct page *hugetlb_nopage(struct vm_area_struct *vma,
unsigned long address, int *unused)
{
BUG();
return NULL;
}
struct vm_operations_struct hugetlb_vm_ops = {
.nopage = hugetlb_nopage,
};
static pte_t make_huge_pte(struct vm_area_struct *vma, struct page *page,
int writable)
{
pte_t entry;
if (writable) {
entry =
pte_mkwrite(pte_mkdirty(mk_pte(page, vma->vm_page_prot)));
} else {
entry = pte_wrprotect(mk_pte(page, vma->vm_page_prot));
}
entry = pte_mkyoung(entry);
entry = pte_mkhuge(entry);
return entry;
}
static void set_huge_ptep_writable(struct vm_area_struct *vma,
unsigned long address, pte_t *ptep)
{
pte_t entry;
entry = pte_mkwrite(pte_mkdirty(*ptep));
ptep_set_access_flags(vma, address, ptep, entry, 1);
update_mmu_cache(vma, address, entry);
lazy_mmu_prot_update(entry);
}
int copy_hugetlb_page_range(struct mm_struct *dst, struct mm_struct *src,
struct vm_area_struct *vma)
{
pte_t *src_pte, *dst_pte, entry;
struct page *ptepage;
unsigned long addr;
int cow;
cow = (vma->vm_flags & (VM_SHARED | VM_MAYWRITE)) == VM_MAYWRITE;
for (addr = vma->vm_start; addr < vma->vm_end; addr += HPAGE_SIZE) {
src_pte = huge_pte_offset(src, addr);
if (!src_pte)
continue;
dst_pte = huge_pte_alloc(dst, addr);
if (!dst_pte)
goto nomem;
spin_lock(&dst->page_table_lock);
spin_lock(&src->page_table_lock);
if (!pte_none(*src_pte)) {
if (cow)
ptep_set_wrprotect(src, addr, src_pte);
entry = *src_pte;
ptepage = pte_page(entry);
get_page(ptepage);
add_mm_counter(dst, file_rss, HPAGE_SIZE / PAGE_SIZE);
set_huge_pte_at(dst, addr, dst_pte, entry);
}
spin_unlock(&src->page_table_lock);
spin_unlock(&dst->page_table_lock);
}
return 0;
nomem:
return -ENOMEM;
}
void unmap_hugepage_range(struct vm_area_struct *vma, unsigned long start,
unsigned long end)
{
struct mm_struct *mm = vma->vm_mm;
unsigned long address;
pte_t *ptep;
pte_t pte;
struct page *page;
WARN_ON(!is_vm_hugetlb_page(vma));
BUG_ON(start & ~HPAGE_MASK);
BUG_ON(end & ~HPAGE_MASK);
spin_lock(&mm->page_table_lock);
[PATCH] mm: update_hiwaters just in time update_mem_hiwater has attracted various criticisms, in particular from those concerned with mm scalability. Originally it was called whenever rss or total_vm got raised. Then many of those callsites were replaced by a timer tick call from account_system_time. Now Frank van Maarseveen reports that to be found inadequate. How about this? Works for Frank. Replace update_mem_hiwater, a poor combination of two unrelated ops, by macros update_hiwater_rss and update_hiwater_vm. Don't attempt to keep mm->hiwater_rss up to date at timer tick, nor every time we raise rss (usually by 1): those are hot paths. Do the opposite, update only when about to lower rss (usually by many), or just before final accounting in do_exit. Handle mm->hiwater_vm in the same way, though it's much less of an issue. Demand that whoever collects these hiwater statistics do the work of taking the maximum with rss or total_vm. And there has been no collector of these hiwater statistics in the tree. The new convention needs an example, so match Frank's usage by adding a VmPeak line above VmSize to /proc/<pid>/status, and also a VmHWM line above VmRSS (High-Water-Mark or High-Water-Memory). There was a particular anomaly during mremap move, that hiwater_vm might be captured too high. A fleeting such anomaly remains, but it's quickly corrected now, whereas before it would stick. What locking? None: if the app is racy then these statistics will be racy, it's not worth any overhead to make them exact. But whenever it suits, hiwater_vm is updated under exclusive mmap_sem, and hiwater_rss under page_table_lock (for now) or with preemption disabled (later on): without going to any trouble, minimize the time between reading current values and updating, to minimize those occasions when a racing thread bumps a count up and back down in between. Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-30 01:16:18 +00:00
/* Update high watermark before we lower rss */
update_hiwater_rss(mm);
for (address = start; address < end; address += HPAGE_SIZE) {
ptep = huge_pte_offset(mm, address);
if (!ptep)
continue;
pte = huge_ptep_get_and_clear(mm, address, ptep);
if (pte_none(pte))
continue;
page = pte_page(pte);
put_page(page);
add_mm_counter(mm, file_rss, (int) -(HPAGE_SIZE / PAGE_SIZE));
}
spin_unlock(&mm->page_table_lock);
flush_tlb_range(vma, start, end);
}
static int hugetlb_cow(struct mm_struct *mm, struct vm_area_struct *vma,
unsigned long address, pte_t *ptep, pte_t pte)
{
struct page *old_page, *new_page;
int avoidcopy;
old_page = pte_page(pte);
/* If no-one else is actually using this page, avoid the copy
* and just make the page writable */
avoidcopy = (page_count(old_page) == 1);
if (avoidcopy) {
set_huge_ptep_writable(vma, address, ptep);
return VM_FAULT_MINOR;
}
page_cache_get(old_page);
new_page = alloc_huge_page(vma, address);
if (!new_page) {
page_cache_release(old_page);
return VM_FAULT_OOM;
}
spin_unlock(&mm->page_table_lock);
copy_huge_page(new_page, old_page, address);
spin_lock(&mm->page_table_lock);
ptep = huge_pte_offset(mm, address & HPAGE_MASK);
if (likely(pte_same(*ptep, pte))) {
/* Break COW */
set_huge_pte_at(mm, address, ptep,
make_huge_pte(vma, new_page, 1));
/* Make the old page be freed below */
new_page = old_page;
}
page_cache_release(new_page);
page_cache_release(old_page);
return VM_FAULT_MINOR;
}
int hugetlb_no_page(struct mm_struct *mm, struct vm_area_struct *vma,
unsigned long address, pte_t *ptep, int write_access)
{
int ret = VM_FAULT_SIGBUS;
unsigned long idx;
unsigned long size;
struct page *page;
struct address_space *mapping;
pte_t new_pte;
mapping = vma->vm_file->f_mapping;
idx = ((address - vma->vm_start) >> HPAGE_SHIFT)
+ (vma->vm_pgoff >> (HPAGE_SHIFT - PAGE_SHIFT));
/*
* Use page lock to guard against racing truncation
* before we get page_table_lock.
*/
retry:
page = find_lock_page(mapping, idx);
if (!page) {
if (hugetlb_get_quota(mapping))
goto out;
page = alloc_huge_page(vma, address);
if (!page) {
hugetlb_put_quota(mapping);
ret = VM_FAULT_OOM;
goto out;
}
clear_huge_page(page, address);
if (vma->vm_flags & VM_SHARED) {
int err;
err = add_to_page_cache(page, mapping, idx, GFP_KERNEL);
if (err) {
put_page(page);
hugetlb_put_quota(mapping);
if (err == -EEXIST)
goto retry;
goto out;
}
} else
lock_page(page);
}
spin_lock(&mm->page_table_lock);
size = i_size_read(mapping->host) >> HPAGE_SHIFT;
if (idx >= size)
goto backout;
ret = VM_FAULT_MINOR;
if (!pte_none(*ptep))
goto backout;
add_mm_counter(mm, file_rss, HPAGE_SIZE / PAGE_SIZE);
new_pte = make_huge_pte(vma, page, ((vma->vm_flags & VM_WRITE)
&& (vma->vm_flags & VM_SHARED)));
set_huge_pte_at(mm, address, ptep, new_pte);
if (write_access && !(vma->vm_flags & VM_SHARED)) {
/* Optimization, do the COW without a second fault */
ret = hugetlb_cow(mm, vma, address, ptep, new_pte);
}
spin_unlock(&mm->page_table_lock);
unlock_page(page);
out:
return ret;
backout:
spin_unlock(&mm->page_table_lock);
hugetlb_put_quota(mapping);
unlock_page(page);
put_page(page);
goto out;
}
int hugetlb_fault(struct mm_struct *mm, struct vm_area_struct *vma,
unsigned long address, int write_access)
{
pte_t *ptep;
pte_t entry;
int ret;
[PATCH] hugepage: serialize hugepage allocation and instantiation Currently, no lock or mutex is held between allocating a hugepage and inserting it into the pagetables / page cache. When we do go to insert the page into pagetables or page cache, we recheck and may free the newly allocated hugepage. However, since the number of hugepages in the system is strictly limited, and it's usualy to want to use all of them, this can still lead to spurious allocation failures. For example, suppose two processes are both mapping (MAP_SHARED) the same hugepage file, large enough to consume the entire available hugepage pool. If they race instantiating the last page in the mapping, they will both attempt to allocate the last available hugepage. One will fail, of course, returning OOM from the fault and thus causing the process to be killed, despite the fact that the entire mapping can, in fact, be instantiated. The patch fixes this race by the simple method of adding a (sleeping) mutex to serialize the hugepage fault path between allocation and insertion into pagetables and/or page cache. It would be possible to avoid the serialization by catching the allocation failures, waiting on some condition, then rechecking to see if someone else has instantiated the page for us. Given the likely frequency of hugepage instantiations, it seems very doubtful it's worth the extra complexity. This patch causes no regression on the libhugetlbfs testsuite, and one test, which can trigger this race now passes where it previously failed. Actually, the test still sometimes fails, though less often and only as a shmat() failure, rather processes getting OOM killed by the VM. The dodgy heuristic tests in fs/hugetlbfs/inode.c for whether there's enough hugepage space aren't protected by the new mutex, and would be ugly to do so, so there's still a race there. Another patch to replace those tests with something saner for this reason as well as others coming... Signed-off-by: David Gibson <dwg@au1.ibm.com> Cc: William Lee Irwin III <wli@holomorphy.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-22 08:08:53 +00:00
static DEFINE_MUTEX(hugetlb_instantiation_mutex);
ptep = huge_pte_alloc(mm, address);
if (!ptep)
return VM_FAULT_OOM;
[PATCH] hugepage: serialize hugepage allocation and instantiation Currently, no lock or mutex is held between allocating a hugepage and inserting it into the pagetables / page cache. When we do go to insert the page into pagetables or page cache, we recheck and may free the newly allocated hugepage. However, since the number of hugepages in the system is strictly limited, and it's usualy to want to use all of them, this can still lead to spurious allocation failures. For example, suppose two processes are both mapping (MAP_SHARED) the same hugepage file, large enough to consume the entire available hugepage pool. If they race instantiating the last page in the mapping, they will both attempt to allocate the last available hugepage. One will fail, of course, returning OOM from the fault and thus causing the process to be killed, despite the fact that the entire mapping can, in fact, be instantiated. The patch fixes this race by the simple method of adding a (sleeping) mutex to serialize the hugepage fault path between allocation and insertion into pagetables and/or page cache. It would be possible to avoid the serialization by catching the allocation failures, waiting on some condition, then rechecking to see if someone else has instantiated the page for us. Given the likely frequency of hugepage instantiations, it seems very doubtful it's worth the extra complexity. This patch causes no regression on the libhugetlbfs testsuite, and one test, which can trigger this race now passes where it previously failed. Actually, the test still sometimes fails, though less often and only as a shmat() failure, rather processes getting OOM killed by the VM. The dodgy heuristic tests in fs/hugetlbfs/inode.c for whether there's enough hugepage space aren't protected by the new mutex, and would be ugly to do so, so there's still a race there. Another patch to replace those tests with something saner for this reason as well as others coming... Signed-off-by: David Gibson <dwg@au1.ibm.com> Cc: William Lee Irwin III <wli@holomorphy.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-22 08:08:53 +00:00
/*
* Serialize hugepage allocation and instantiation, so that we don't
* get spurious allocation failures if two CPUs race to instantiate
* the same page in the page cache.
*/
mutex_lock(&hugetlb_instantiation_mutex);
entry = *ptep;
[PATCH] hugepage: serialize hugepage allocation and instantiation Currently, no lock or mutex is held between allocating a hugepage and inserting it into the pagetables / page cache. When we do go to insert the page into pagetables or page cache, we recheck and may free the newly allocated hugepage. However, since the number of hugepages in the system is strictly limited, and it's usualy to want to use all of them, this can still lead to spurious allocation failures. For example, suppose two processes are both mapping (MAP_SHARED) the same hugepage file, large enough to consume the entire available hugepage pool. If they race instantiating the last page in the mapping, they will both attempt to allocate the last available hugepage. One will fail, of course, returning OOM from the fault and thus causing the process to be killed, despite the fact that the entire mapping can, in fact, be instantiated. The patch fixes this race by the simple method of adding a (sleeping) mutex to serialize the hugepage fault path between allocation and insertion into pagetables and/or page cache. It would be possible to avoid the serialization by catching the allocation failures, waiting on some condition, then rechecking to see if someone else has instantiated the page for us. Given the likely frequency of hugepage instantiations, it seems very doubtful it's worth the extra complexity. This patch causes no regression on the libhugetlbfs testsuite, and one test, which can trigger this race now passes where it previously failed. Actually, the test still sometimes fails, though less often and only as a shmat() failure, rather processes getting OOM killed by the VM. The dodgy heuristic tests in fs/hugetlbfs/inode.c for whether there's enough hugepage space aren't protected by the new mutex, and would be ugly to do so, so there's still a race there. Another patch to replace those tests with something saner for this reason as well as others coming... Signed-off-by: David Gibson <dwg@au1.ibm.com> Cc: William Lee Irwin III <wli@holomorphy.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-22 08:08:53 +00:00
if (pte_none(entry)) {
ret = hugetlb_no_page(mm, vma, address, ptep, write_access);
mutex_unlock(&hugetlb_instantiation_mutex);
return ret;
}
ret = VM_FAULT_MINOR;
spin_lock(&mm->page_table_lock);
/* Check for a racing update before calling hugetlb_cow */
if (likely(pte_same(entry, *ptep)))
if (write_access && !pte_write(entry))
ret = hugetlb_cow(mm, vma, address, ptep, entry);
spin_unlock(&mm->page_table_lock);
[PATCH] hugepage: serialize hugepage allocation and instantiation Currently, no lock or mutex is held between allocating a hugepage and inserting it into the pagetables / page cache. When we do go to insert the page into pagetables or page cache, we recheck and may free the newly allocated hugepage. However, since the number of hugepages in the system is strictly limited, and it's usualy to want to use all of them, this can still lead to spurious allocation failures. For example, suppose two processes are both mapping (MAP_SHARED) the same hugepage file, large enough to consume the entire available hugepage pool. If they race instantiating the last page in the mapping, they will both attempt to allocate the last available hugepage. One will fail, of course, returning OOM from the fault and thus causing the process to be killed, despite the fact that the entire mapping can, in fact, be instantiated. The patch fixes this race by the simple method of adding a (sleeping) mutex to serialize the hugepage fault path between allocation and insertion into pagetables and/or page cache. It would be possible to avoid the serialization by catching the allocation failures, waiting on some condition, then rechecking to see if someone else has instantiated the page for us. Given the likely frequency of hugepage instantiations, it seems very doubtful it's worth the extra complexity. This patch causes no regression on the libhugetlbfs testsuite, and one test, which can trigger this race now passes where it previously failed. Actually, the test still sometimes fails, though less often and only as a shmat() failure, rather processes getting OOM killed by the VM. The dodgy heuristic tests in fs/hugetlbfs/inode.c for whether there's enough hugepage space aren't protected by the new mutex, and would be ugly to do so, so there's still a race there. Another patch to replace those tests with something saner for this reason as well as others coming... Signed-off-by: David Gibson <dwg@au1.ibm.com> Cc: William Lee Irwin III <wli@holomorphy.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-22 08:08:53 +00:00
mutex_unlock(&hugetlb_instantiation_mutex);
return ret;
}
int follow_hugetlb_page(struct mm_struct *mm, struct vm_area_struct *vma,
struct page **pages, struct vm_area_struct **vmas,
unsigned long *position, int *length, int i)
{
unsigned long pfn_offset;
unsigned long vaddr = *position;
int remainder = *length;
spin_lock(&mm->page_table_lock);
while (vaddr < vma->vm_end && remainder) {
pte_t *pte;
struct page *page;
/*
* Some archs (sparc64, sh*) have multiple pte_ts to
* each hugepage. We have to make * sure we get the
* first, for the page indexing below to work.
*/
pte = huge_pte_offset(mm, vaddr & HPAGE_MASK);
if (!pte || pte_none(*pte)) {
int ret;
spin_unlock(&mm->page_table_lock);
ret = hugetlb_fault(mm, vma, vaddr, 0);
spin_lock(&mm->page_table_lock);
if (ret == VM_FAULT_MINOR)
continue;
remainder = 0;
if (!i)
i = -EFAULT;
break;
}
pfn_offset = (vaddr & ~HPAGE_MASK) >> PAGE_SHIFT;
page = pte_page(*pte);
same_page:
if (pages) {
get_page(page);
pages[i] = page + pfn_offset;
}
if (vmas)
vmas[i] = vma;
vaddr += PAGE_SIZE;
++pfn_offset;
--remainder;
++i;
if (vaddr < vma->vm_end && remainder &&
pfn_offset < HPAGE_SIZE/PAGE_SIZE) {
/*
* We use pfn_offset to avoid touching the pageframes
* of this compound page.
*/
goto same_page;
}
}
spin_unlock(&mm->page_table_lock);
*length = remainder;
*position = vaddr;
return i;
}
[PATCH] Enable mprotect on huge pages 2.6.16-rc3 uses hugetlb on-demand paging, but it doesn_t support hugetlb mprotect. From: David Gibson <david@gibson.dropbear.id.au> Remove a test from the mprotect() path which checks that the mprotect()ed range on a hugepage VMA is hugepage aligned (yes, really, the sense of is_aligned_hugepage_range() is the opposite of what you'd guess :-/). In fact, we don't need this test. If the given addresses match the beginning/end of a hugepage VMA they must already be suitably aligned. If they don't, then mprotect_fixup() will attempt to split the VMA. The very first test in split_vma() will check for a badly aligned address on a hugepage VMA and return -EINVAL if necessary. From: "Chen, Kenneth W" <kenneth.w.chen@intel.com> On i386 and x86-64, pte flag _PAGE_PSE collides with _PAGE_PROTNONE. The identify of hugetlb pte is lost when changing page protection via mprotect. A page fault occurs later will trigger a bug check in huge_pte_alloc(). The fix is to always make new pte a hugetlb pte and also to clean up legacy code where _PAGE_PRESENT is forced on in the pre-faulting day. Signed-off-by: Zhang Yanmin <yanmin.zhang@intel.com> Cc: David Gibson <david@gibson.dropbear.id.au> Cc: "David S. Miller" <davem@davemloft.net> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Paul Mackerras <paulus@samba.org> Cc: William Lee Irwin III <wli@holomorphy.com> Signed-off-by: Ken Chen <kenneth.w.chen@intel.com> Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com> Cc: Andi Kleen <ak@muc.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-03-22 08:08:50 +00:00
void hugetlb_change_protection(struct vm_area_struct *vma,
unsigned long address, unsigned long end, pgprot_t newprot)
{
struct mm_struct *mm = vma->vm_mm;
unsigned long start = address;
pte_t *ptep;
pte_t pte;
BUG_ON(address >= end);
flush_cache_range(vma, address, end);
spin_lock(&mm->page_table_lock);
for (; address < end; address += HPAGE_SIZE) {
ptep = huge_pte_offset(mm, address);
if (!ptep)
continue;
if (!pte_none(*ptep)) {
pte = huge_ptep_get_and_clear(mm, address, ptep);
pte = pte_mkhuge(pte_modify(pte, newprot));
set_huge_pte_at(mm, address, ptep, pte);
lazy_mmu_prot_update(pte);
}
}
spin_unlock(&mm->page_table_lock);
flush_tlb_range(vma, start, end);
}
struct file_region {
struct list_head link;
long from;
long to;
};
static long region_add(struct list_head *head, long f, long t)
{
struct file_region *rg, *nrg, *trg;
/* Locate the region we are either in or before. */
list_for_each_entry(rg, head, link)
if (f <= rg->to)
break;
/* Round our left edge to the current segment if it encloses us. */
if (f > rg->from)
f = rg->from;
/* Check for and consume any regions we now overlap with. */
nrg = rg;
list_for_each_entry_safe(rg, trg, rg->link.prev, link) {
if (&rg->link == head)
break;
if (rg->from > t)
break;
/* If this area reaches higher then extend our area to
* include it completely. If this is not the first area
* which we intend to reuse, free it. */
if (rg->to > t)
t = rg->to;
if (rg != nrg) {
list_del(&rg->link);
kfree(rg);
}
}
nrg->from = f;
nrg->to = t;
return 0;
}
static long region_chg(struct list_head *head, long f, long t)
{
struct file_region *rg, *nrg;
long chg = 0;
/* Locate the region we are before or in. */
list_for_each_entry(rg, head, link)
if (f <= rg->to)
break;
/* If we are below the current region then a new region is required.
* Subtle, allocate a new region at the position but make it zero
* size such that we can guarentee to record the reservation. */
if (&rg->link == head || t < rg->from) {
nrg = kmalloc(sizeof(*nrg), GFP_KERNEL);
if (nrg == 0)
return -ENOMEM;
nrg->from = f;
nrg->to = f;
INIT_LIST_HEAD(&nrg->link);
list_add(&nrg->link, rg->link.prev);
return t - f;
}
/* Round our left edge to the current segment if it encloses us. */
if (f > rg->from)
f = rg->from;
chg = t - f;
/* Check for and consume any regions we now overlap with. */
list_for_each_entry(rg, rg->link.prev, link) {
if (&rg->link == head)
break;
if (rg->from > t)
return chg;
/* We overlap with this area, if it extends futher than
* us then we must extend ourselves. Account for its
* existing reservation. */
if (rg->to > t) {
chg += rg->to - t;
t = rg->to;
}
chg -= rg->to - rg->from;
}
return chg;
}
static long region_truncate(struct list_head *head, long end)
{
struct file_region *rg, *trg;
long chg = 0;
/* Locate the region we are either in or before. */
list_for_each_entry(rg, head, link)
if (end <= rg->to)
break;
if (&rg->link == head)
return 0;
/* If we are in the middle of a region then adjust it. */
if (end > rg->from) {
chg = rg->to - end;
rg->to = end;
rg = list_entry(rg->link.next, typeof(*rg), link);
}
/* Drop any remaining regions. */
list_for_each_entry_safe(rg, trg, rg->link.prev, link) {
if (&rg->link == head)
break;
chg += rg->to - rg->from;
list_del(&rg->link);
kfree(rg);
}
return chg;
}
static int hugetlb_acct_memory(long delta)
{
int ret = -ENOMEM;
spin_lock(&hugetlb_lock);
if ((delta + resv_huge_pages) <= free_huge_pages) {
resv_huge_pages += delta;
ret = 0;
}
spin_unlock(&hugetlb_lock);
return ret;
}
int hugetlb_reserve_pages(struct inode *inode, long from, long to)
{
long ret, chg;
chg = region_chg(&inode->i_mapping->private_list, from, to);
if (chg < 0)
return chg;
ret = hugetlb_acct_memory(chg);
if (ret < 0)
return ret;
region_add(&inode->i_mapping->private_list, from, to);
return 0;
}
void hugetlb_unreserve_pages(struct inode *inode, long offset, long freed)
{
long chg = region_truncate(&inode->i_mapping->private_list, offset);
hugetlb_acct_memory(freed - chg);
}