linux-stable/arch/arm/mm/mmap.c

323 lines
7.8 KiB
C
Raw Normal View History

/*
* linux/arch/arm/mm/mmap.c
*/
#include <linux/fs.h>
#include <linux/mm.h>
#include <linux/mman.h>
#include <linux/shm.h>
#include <linux/sched.h>
#include <linux/io.h>
#include <linux/personality.h>
#include <linux/random.h>
#include <asm/cachetype.h>
static inline unsigned long COLOUR_ALIGN_DOWN(unsigned long addr,
unsigned long pgoff)
{
unsigned long base = addr & ~(SHMLBA-1);
unsigned long off = (pgoff << PAGE_SHIFT) & (SHMLBA-1);
if (base + off <= addr)
return base + off;
return base - off;
}
#define COLOUR_ALIGN(addr,pgoff) \
((((addr)+SHMLBA-1)&~(SHMLBA-1)) + \
(((pgoff)<<PAGE_SHIFT) & (SHMLBA-1)))
/* gap between mmap and stack */
#define MIN_GAP (128*1024*1024UL)
#define MAX_GAP ((TASK_SIZE)/6*5)
static int mmap_is_legacy(void)
{
if (current->personality & ADDR_COMPAT_LAYOUT)
return 1;
if (rlimit(RLIMIT_STACK) == RLIM_INFINITY)
return 1;
return sysctl_legacy_va_layout;
}
static unsigned long mmap_base(unsigned long rnd)
{
unsigned long gap = rlimit(RLIMIT_STACK);
if (gap < MIN_GAP)
gap = MIN_GAP;
else if (gap > MAX_GAP)
gap = MAX_GAP;
return PAGE_ALIGN(TASK_SIZE - gap - rnd);
}
/*
* We need to ensure that shared mappings are correctly aligned to
* avoid aliasing issues with VIPT caches. We need to ensure that
* a specific page of an object is always mapped at a multiple of
* SHMLBA bytes.
*
* We unconditionally provide this function for all cases, however
* in the VIVT case, we optimise out the alignment rules.
*/
unsigned long
arch_get_unmapped_area(struct file *filp, unsigned long addr,
unsigned long len, unsigned long pgoff, unsigned long flags)
{
struct mm_struct *mm = current->mm;
struct vm_area_struct *vma;
unsigned long start_addr;
int do_align = 0;
int aliasing = cache_is_vipt_aliasing();
/*
* We only need to do colour alignment if either the I or D
* caches alias.
*/
if (aliasing)
do_align = filp || (flags & MAP_SHARED);
/*
* We enforce the MAP_FIXED case.
*/
if (flags & MAP_FIXED) {
if (aliasing && flags & MAP_SHARED &&
(addr - (pgoff << PAGE_SHIFT)) & (SHMLBA - 1))
return -EINVAL;
return addr;
}
if (len > TASK_SIZE)
return -ENOMEM;
if (addr) {
if (do_align)
addr = COLOUR_ALIGN(addr, pgoff);
else
addr = PAGE_ALIGN(addr);
vma = find_vma(mm, addr);
if (TASK_SIZE - len >= addr &&
(!vma || addr + len <= vma->vm_start))
return addr;
}
[PATCH] Avoiding mmap fragmentation Ingo recently introduced a great speedup for allocating new mmaps using the free_area_cache pointer which boosts the specweb SSL benchmark by 4-5% and causes huge performance increases in thread creation. The downside of this patch is that it does lead to fragmentation in the mmap-ed areas (visible via /proc/self/maps), such that some applications that work fine under 2.4 kernels quickly run out of memory on any 2.6 kernel. The problem is twofold: 1) the free_area_cache is used to continue a search for memory where the last search ended. Before the change new areas were always searched from the base address on. So now new small areas are cluttering holes of all sizes throughout the whole mmap-able region whereas before small holes tended to close holes near the base leaving holes far from the base large and available for larger requests. 2) the free_area_cache also is set to the location of the last munmap-ed area so in scenarios where we allocate e.g. five regions of 1K each, then free regions 4 2 3 in this order the next request for 1K will be placed in the position of the old region 3, whereas before we appended it to the still active region 1, placing it at the location of the old region 2. Before we had 1 free region of 2K, now we only get two free regions of 1K -> fragmentation. The patch addresses thes issues by introducing yet another cache descriptor cached_hole_size that contains the largest known hole size below the current free_area_cache. If a new request comes in the size is compared against the cached_hole_size and if the request can be filled with a hole below free_area_cache the search is started from the base instead. The results look promising: Whereas 2.6.12-rc4 fragments quickly and my (earlier posted) leakme.c test program terminates after 50000+ iterations with 96 distinct and fragmented maps in /proc/self/maps it performs nicely (as expected) with thread creation, Ingo's test_str02 with 20000 threads requires 0.7s system time. Taking out Ingo's patch (un-patch available per request) by basically deleting all mentions of free_area_cache from the kernel and starting the search for new memory always at the respective bases we observe: leakme terminates successfully with 11 distinctive hardly fragmented areas in /proc/self/maps but thread creating is gringdingly slow: 30+s(!) system time for Ingo's test_str02 with 20000 threads. Now - drumroll ;-) the appended patch works fine with leakme: it ends with only 7 distinct areas in /proc/self/maps and also thread creation seems sufficiently fast with 0.71s for 20000 threads. Signed-off-by: Wolfgang Wander <wwc@rentec.com> Credit-to: "Richard Purdie" <rpurdie@rpsys.net> Signed-off-by: Ken Chen <kenneth.w.chen@intel.com> Acked-by: Ingo Molnar <mingo@elte.hu> (partly) Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-22 00:14:49 +00:00
if (len > mm->cached_hole_size) {
start_addr = addr = mm->free_area_cache;
} else {
start_addr = addr = mm->mmap_base;
[PATCH] Avoiding mmap fragmentation Ingo recently introduced a great speedup for allocating new mmaps using the free_area_cache pointer which boosts the specweb SSL benchmark by 4-5% and causes huge performance increases in thread creation. The downside of this patch is that it does lead to fragmentation in the mmap-ed areas (visible via /proc/self/maps), such that some applications that work fine under 2.4 kernels quickly run out of memory on any 2.6 kernel. The problem is twofold: 1) the free_area_cache is used to continue a search for memory where the last search ended. Before the change new areas were always searched from the base address on. So now new small areas are cluttering holes of all sizes throughout the whole mmap-able region whereas before small holes tended to close holes near the base leaving holes far from the base large and available for larger requests. 2) the free_area_cache also is set to the location of the last munmap-ed area so in scenarios where we allocate e.g. five regions of 1K each, then free regions 4 2 3 in this order the next request for 1K will be placed in the position of the old region 3, whereas before we appended it to the still active region 1, placing it at the location of the old region 2. Before we had 1 free region of 2K, now we only get two free regions of 1K -> fragmentation. The patch addresses thes issues by introducing yet another cache descriptor cached_hole_size that contains the largest known hole size below the current free_area_cache. If a new request comes in the size is compared against the cached_hole_size and if the request can be filled with a hole below free_area_cache the search is started from the base instead. The results look promising: Whereas 2.6.12-rc4 fragments quickly and my (earlier posted) leakme.c test program terminates after 50000+ iterations with 96 distinct and fragmented maps in /proc/self/maps it performs nicely (as expected) with thread creation, Ingo's test_str02 with 20000 threads requires 0.7s system time. Taking out Ingo's patch (un-patch available per request) by basically deleting all mentions of free_area_cache from the kernel and starting the search for new memory always at the respective bases we observe: leakme terminates successfully with 11 distinctive hardly fragmented areas in /proc/self/maps but thread creating is gringdingly slow: 30+s(!) system time for Ingo's test_str02 with 20000 threads. Now - drumroll ;-) the appended patch works fine with leakme: it ends with only 7 distinct areas in /proc/self/maps and also thread creation seems sufficiently fast with 0.71s for 20000 threads. Signed-off-by: Wolfgang Wander <wwc@rentec.com> Credit-to: "Richard Purdie" <rpurdie@rpsys.net> Signed-off-by: Ken Chen <kenneth.w.chen@intel.com> Acked-by: Ingo Molnar <mingo@elte.hu> (partly) Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-22 00:14:49 +00:00
mm->cached_hole_size = 0;
}
full_search:
if (do_align)
addr = COLOUR_ALIGN(addr, pgoff);
else
addr = PAGE_ALIGN(addr);
for (vma = find_vma(mm, addr); ; vma = vma->vm_next) {
/* At this point: (!vma || addr < vma->vm_end). */
if (TASK_SIZE - len < addr) {
/*
* Start a new search - just in case we missed
* some holes.
*/
if (start_addr != TASK_UNMAPPED_BASE) {
start_addr = addr = TASK_UNMAPPED_BASE;
[PATCH] Avoiding mmap fragmentation Ingo recently introduced a great speedup for allocating new mmaps using the free_area_cache pointer which boosts the specweb SSL benchmark by 4-5% and causes huge performance increases in thread creation. The downside of this patch is that it does lead to fragmentation in the mmap-ed areas (visible via /proc/self/maps), such that some applications that work fine under 2.4 kernels quickly run out of memory on any 2.6 kernel. The problem is twofold: 1) the free_area_cache is used to continue a search for memory where the last search ended. Before the change new areas were always searched from the base address on. So now new small areas are cluttering holes of all sizes throughout the whole mmap-able region whereas before small holes tended to close holes near the base leaving holes far from the base large and available for larger requests. 2) the free_area_cache also is set to the location of the last munmap-ed area so in scenarios where we allocate e.g. five regions of 1K each, then free regions 4 2 3 in this order the next request for 1K will be placed in the position of the old region 3, whereas before we appended it to the still active region 1, placing it at the location of the old region 2. Before we had 1 free region of 2K, now we only get two free regions of 1K -> fragmentation. The patch addresses thes issues by introducing yet another cache descriptor cached_hole_size that contains the largest known hole size below the current free_area_cache. If a new request comes in the size is compared against the cached_hole_size and if the request can be filled with a hole below free_area_cache the search is started from the base instead. The results look promising: Whereas 2.6.12-rc4 fragments quickly and my (earlier posted) leakme.c test program terminates after 50000+ iterations with 96 distinct and fragmented maps in /proc/self/maps it performs nicely (as expected) with thread creation, Ingo's test_str02 with 20000 threads requires 0.7s system time. Taking out Ingo's patch (un-patch available per request) by basically deleting all mentions of free_area_cache from the kernel and starting the search for new memory always at the respective bases we observe: leakme terminates successfully with 11 distinctive hardly fragmented areas in /proc/self/maps but thread creating is gringdingly slow: 30+s(!) system time for Ingo's test_str02 with 20000 threads. Now - drumroll ;-) the appended patch works fine with leakme: it ends with only 7 distinct areas in /proc/self/maps and also thread creation seems sufficiently fast with 0.71s for 20000 threads. Signed-off-by: Wolfgang Wander <wwc@rentec.com> Credit-to: "Richard Purdie" <rpurdie@rpsys.net> Signed-off-by: Ken Chen <kenneth.w.chen@intel.com> Acked-by: Ingo Molnar <mingo@elte.hu> (partly) Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-22 00:14:49 +00:00
mm->cached_hole_size = 0;
goto full_search;
}
return -ENOMEM;
}
if (!vma || addr + len <= vma->vm_start) {
/*
* Remember the place where we stopped the search:
*/
mm->free_area_cache = addr + len;
return addr;
}
[PATCH] Avoiding mmap fragmentation Ingo recently introduced a great speedup for allocating new mmaps using the free_area_cache pointer which boosts the specweb SSL benchmark by 4-5% and causes huge performance increases in thread creation. The downside of this patch is that it does lead to fragmentation in the mmap-ed areas (visible via /proc/self/maps), such that some applications that work fine under 2.4 kernels quickly run out of memory on any 2.6 kernel. The problem is twofold: 1) the free_area_cache is used to continue a search for memory where the last search ended. Before the change new areas were always searched from the base address on. So now new small areas are cluttering holes of all sizes throughout the whole mmap-able region whereas before small holes tended to close holes near the base leaving holes far from the base large and available for larger requests. 2) the free_area_cache also is set to the location of the last munmap-ed area so in scenarios where we allocate e.g. five regions of 1K each, then free regions 4 2 3 in this order the next request for 1K will be placed in the position of the old region 3, whereas before we appended it to the still active region 1, placing it at the location of the old region 2. Before we had 1 free region of 2K, now we only get two free regions of 1K -> fragmentation. The patch addresses thes issues by introducing yet another cache descriptor cached_hole_size that contains the largest known hole size below the current free_area_cache. If a new request comes in the size is compared against the cached_hole_size and if the request can be filled with a hole below free_area_cache the search is started from the base instead. The results look promising: Whereas 2.6.12-rc4 fragments quickly and my (earlier posted) leakme.c test program terminates after 50000+ iterations with 96 distinct and fragmented maps in /proc/self/maps it performs nicely (as expected) with thread creation, Ingo's test_str02 with 20000 threads requires 0.7s system time. Taking out Ingo's patch (un-patch available per request) by basically deleting all mentions of free_area_cache from the kernel and starting the search for new memory always at the respective bases we observe: leakme terminates successfully with 11 distinctive hardly fragmented areas in /proc/self/maps but thread creating is gringdingly slow: 30+s(!) system time for Ingo's test_str02 with 20000 threads. Now - drumroll ;-) the appended patch works fine with leakme: it ends with only 7 distinct areas in /proc/self/maps and also thread creation seems sufficiently fast with 0.71s for 20000 threads. Signed-off-by: Wolfgang Wander <wwc@rentec.com> Credit-to: "Richard Purdie" <rpurdie@rpsys.net> Signed-off-by: Ken Chen <kenneth.w.chen@intel.com> Acked-by: Ingo Molnar <mingo@elte.hu> (partly) Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-06-22 00:14:49 +00:00
if (addr + mm->cached_hole_size < vma->vm_start)
mm->cached_hole_size = vma->vm_start - addr;
addr = vma->vm_end;
if (do_align)
addr = COLOUR_ALIGN(addr, pgoff);
}
}
unsigned long
arch_get_unmapped_area_topdown(struct file *filp, const unsigned long addr0,
const unsigned long len, const unsigned long pgoff,
const unsigned long flags)
{
struct vm_area_struct *vma;
struct mm_struct *mm = current->mm;
unsigned long addr = addr0;
int do_align = 0;
int aliasing = cache_is_vipt_aliasing();
/*
* We only need to do colour alignment if either the I or D
* caches alias.
*/
if (aliasing)
do_align = filp || (flags & MAP_SHARED);
/* requested length too big for entire address space */
if (len > TASK_SIZE)
return -ENOMEM;
if (flags & MAP_FIXED) {
if (aliasing && flags & MAP_SHARED &&
(addr - (pgoff << PAGE_SHIFT)) & (SHMLBA - 1))
return -EINVAL;
return addr;
}
/* requesting a specific address */
if (addr) {
if (do_align)
addr = COLOUR_ALIGN(addr, pgoff);
else
addr = PAGE_ALIGN(addr);
vma = find_vma(mm, addr);
if (TASK_SIZE - len >= addr &&
(!vma || addr + len <= vma->vm_start))
return addr;
}
/* check if free_area_cache is useful for us */
if (len <= mm->cached_hole_size) {
mm->cached_hole_size = 0;
mm->free_area_cache = mm->mmap_base;
}
/* either no address requested or can't fit in requested address hole */
addr = mm->free_area_cache;
if (do_align) {
unsigned long base = COLOUR_ALIGN_DOWN(addr - len, pgoff);
addr = base + len;
}
/* make sure it can fit in the remaining address space */
if (addr > len) {
vma = find_vma(mm, addr-len);
if (!vma || addr <= vma->vm_start)
/* remember the address as a hint for next time */
return (mm->free_area_cache = addr-len);
}
if (mm->mmap_base < len)
goto bottomup;
addr = mm->mmap_base - len;
if (do_align)
addr = COLOUR_ALIGN_DOWN(addr, pgoff);
do {
/*
* Lookup failure means no vma is above this address,
* else if new region fits below vma->vm_start,
* return with success:
*/
vma = find_vma(mm, addr);
if (!vma || addr+len <= vma->vm_start)
/* remember the address as a hint for next time */
return (mm->free_area_cache = addr);
/* remember the largest hole we saw so far */
if (addr + mm->cached_hole_size < vma->vm_start)
mm->cached_hole_size = vma->vm_start - addr;
/* try just below the current vma->vm_start */
addr = vma->vm_start - len;
if (do_align)
addr = COLOUR_ALIGN_DOWN(addr, pgoff);
} while (len < vma->vm_start);
bottomup:
/*
* A failed mmap() very likely causes application failure,
* so fall back to the bottom-up function here. This scenario
* can happen with large stack limits and large mmap()
* allocations.
*/
mm->cached_hole_size = ~0UL;
mm->free_area_cache = TASK_UNMAPPED_BASE;
addr = arch_get_unmapped_area(filp, addr0, len, pgoff, flags);
/*
* Restore the topdown base:
*/
mm->free_area_cache = mm->mmap_base;
mm->cached_hole_size = ~0UL;
return addr;
}
void arch_pick_mmap_layout(struct mm_struct *mm)
{
unsigned long random_factor = 0UL;
/* 8 bits of randomness in 20 address space bits */
if ((current->flags & PF_RANDOMIZE) &&
!(current->personality & ADDR_NO_RANDOMIZE))
random_factor = (get_random_int() % (1 << 8)) << PAGE_SHIFT;
if (mmap_is_legacy()) {
mm->mmap_base = TASK_UNMAPPED_BASE + random_factor;
mm->get_unmapped_area = arch_get_unmapped_area;
mm->unmap_area = arch_unmap_area;
} else {
mm->mmap_base = mmap_base(random_factor);
mm->get_unmapped_area = arch_get_unmapped_area_topdown;
mm->unmap_area = arch_unmap_area_topdown;
}
}
/*
* You really shouldn't be using read() or write() on /dev/mem. This
* might go away in the future.
*/
int valid_phys_addr_range(unsigned long addr, size_t size)
{
if (addr < PHYS_OFFSET)
return 0;
if (addr + size > __pa(high_memory - 1) + 1)
return 0;
return 1;
}
/*
* We don't use supersection mappings for mmap() on /dev/mem, which
* means that we can't map the memory area above the 4G barrier into
* userspace.
*/
int valid_mmap_phys_addr_range(unsigned long pfn, size_t size)
{
return !(pfn + (size >> PAGE_SHIFT) > 0x00100000);
}
#ifdef CONFIG_STRICT_DEVMEM
#include <linux/ioport.h>
/*
* devmem_is_allowed() checks to see if /dev/mem access to a certain
* address is valid. The argument is a physical page number.
* We mimic x86 here by disallowing access to system RAM as well as
* device-exclusive MMIO regions. This effectively disable read()/write()
* on /dev/mem.
*/
int devmem_is_allowed(unsigned long pfn)
{
if (iomem_is_exclusive(pfn << PAGE_SHIFT))
return 0;
if (!page_is_ram(pfn))
return 1;
return 0;
}
#endif