NFS can handle the case where invalidate_inode_pages2_range() fails, so the
premise behind commit 8258d4a574 is now gone.
Remove the WARN_ON_ONCE() which is causing users grief as we can see from
http://bugzilla.kernel.org/show_bug.cgi?id=7826
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This patch fixes core dumps to include the vDSO vma, which is left out now.
It removes the special-case core writing macros, which were not doing the
right thing for the vDSO vma anyway. Instead, it uses VM_ALWAYSDUMP in the
vma; there is no need for the fixmap page to be installed. It handles the
CONFIG_COMPAT_VDSO case by making elf_core_dump use the fake vma from
get_gate_vma after real vmas in the same way the /proc/PID/maps code does.
This changes core dumps so they no longer include the non-PT_LOAD phdrs from
the vDSO. I made the change to add them in the first place, but in turned out
that nothing ever wanted them there since the advent of NT_AUXV. It's cleaner
to leave them out, and just let the phdrs inside the vDSO image speak for
themselves.
Signed-off-by: Roland McGrath <roland@redhat.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This patch fixes the initialization of gate_vma.vm_flags and
gate_vma.vm_page_prot to reflect reality. This makes the "[vdso]" line in
/proc/PID/maps correctly show r-xp instead of ---p, when gate_vma is used
(CONFIG_COMPAT_VDSO on i386).
Signed-off-by: Roland McGrath <roland@redhat.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
It's not pretty, but it appears that ext3 with data=journal will clean
pages without ever actually telling the VM that they are clean. This,
in turn, will result in the VM (and balance_dirty_pages() in particular)
to never realize that the pages got cleaned, and wait forever for an
event that already happened.
Technically, this seems to be a problem with ext3 itself, but it used to
be hidden by 'try_to_free_buffers()' noticing this situation on its own,
and just working around the filesystem problem.
This commit re-instates that hack, in order to avoid a regression for
the 2.6.20 release. This fixes bugzilla 7844:
http://bugzilla.kernel.org/show_bug.cgi?id=7844
Peter Zijlstra points out that we should probably retain the debugging
code that this removes from cancel_dirty_page(), and I agree, but for
the imminent release we might as well just silence the warning too
(since it's not a new bug: anything that triggers that warning has been
around forever).
Acked-by: Randy Dunlap <rdunlap@xenotime.net>
Acked-by: Jens Axboe <jens.axboe@oracle.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Currently one can specify an arbitrary node mask to mbind that includes
nodes not allowed. If that is done with an interleave policy then we will
go around all the nodes. Those outside of the currently allowed cpuset
will be redirected to the border nodes. Interleave will then create
imbalances at the borders of the cpuset.
This patch restricts the nodes to the currently allowed cpuset.
The RFC for this patch was discussed at
http://marc.theaimsgroup.com/?t=116793842100004&r=1&w=2
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Cc: Paul Jackson <pj@sgi.com>
Cc: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Currently we issue a bounce trace when __blk_queue_bounce() is called,
but that merely means that the device has a lower dma mask than the
higher pages in the system. The bio itself may still be lower pages. So
move the bounce trace into __blk_queue_bounce(), when we know there will
actually be page bouncing.
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
NFS: Fix race in nfs_release_page()
invalidate_inode_pages2() may find the dirty bit has been set on a page
owing to the fact that the page may still be mapped after it was locked.
Only after the call to unmap_mapping_range() are we sure that the page
can no longer be dirtied.
In order to fix this, NFS has hooked the releasepage() method and tries
to write the page out between the call to unmap_mapping_range() and the
call to remove_mapping(). This, however leads to deadlocks in the page
reclaim code, where the page may be locked without holding a reference
to the inode or dentry.
Fix is to add a new address_space_operation, launder_page(), which will
attempt to write out a dirty page without releasing the page lock.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Also, the bare SetPageDirty() can skew all sort of accounting leading to
other nasties.
[akpm@osdl.org: cleanup]
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Fix an oops experienced on the Cell architecture when init-time functions,
early_*(), are called at runtime. It alters the call paths to make sure
that the callers explicitly say whether the call is being made on behalf of
a hotplug even, or happening at boot-time.
It has been compile tested on ppc64, ia64, s390, i386 and x86_64.
Acked-by: Arnd Bergmann <arndb@de.ibm.com>
Signed-off-by: Dave Hansen <haveblue@us.ibm.com>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
Acked-by: Andy Whitcroft <apw@shadowen.org>
Cc: Christoph Lameter <clameter@engr.sgi.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Since get_user_pages() may be used with processes other than the
current process and calls flush_anon_page(), flush_anon_page() has to
cope in some way with non-current processes.
It may not be appropriate, or even desirable to flush a region of
virtual memory cache in the current process when that is different to
the process that we want the flush to occur for.
Therefore, pass the vma into flush_anon_page() so that the architecture
can work out whether the 'vmaddr' is for the current process or not.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
At the end of shrink_all_memory() we forget to recalculate lru_pages: it can
be zero.
Fix that up, and add a helper function for this operation too.
Also, recalculate lru_pages each time around the inner loop to get the
balancing correct.
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
These days, if you swapoff when there isn't enough memory, OOM killer gives
"BUG: scheduling while atomic" and the machine hangs: badness() needs to do
its PF_SWAPOFF return after the task_unlock (tasklist_lock is also held
here, so p isn't going to be freed: PF_SWAPOFF might get turned off at any
moment, but that doesn't really matter).
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Both process_zones() and drain_node_pages() check for populated zones
before touching pagesets. However, __drain_pages does not do so,
This may result in a NULL pointer dereference for pagesets in unpopulated
zones if a NUMA setup is combined with cpu hotplug.
Initially the unpopulated zone has the pcp pointers pointing to the boot
pagesets. Since the zone is not populated the boot pageset pointers will
not be changed during page allocator and slab bootstrap.
If a cpu is later brought down (first call to __drain_pages()) then the pcp
pointers for cpus in unpopulated zones are set to NULL since __drain_pages
does not first check for an unpopulated zone.
If the cpu is then brought up again then we call process_zones() which will
ignore the unpopulated zone. So the pageset pointers will still be NULL.
If the cpu is then again brought down then __drain_pages will attempt to
drain pages by following the NULL pageset pointer for unpopulated zones.
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
pdflush hit the BUG_ON(!PageSlab(page)) in kmem_freepages called from
fallback_alloc: cache_grow already freed those pages when alloc_slabmgmt
failed. But it wouldn't have freed them if __GFP_NO_GROW, so make sure
fallback_alloc doesn't waste its time on that case.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Acked-by: Christoph Lameter <clameter@sgi.com>
Acked-by: Pekka J Enberg <penberg@cs.helsinki.fi>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
At the moment the inode/dentry cache hash tables (common by way of
alloc_large_system_hash()) are incorrectly sized by their respective
detection logic when we attempt to use large base pages on systems with
little memory.
This results in odd behaviour when using a 64kB PAGE_SIZE, such as:
Dentry cache hash table entries: 8192 (order: -1, 32768 bytes)
Inode-cache hash table entries: 4096 (order: -2, 16384 bytes)
The mount cache hash table is seemingly the only one that gets this right
by directly taking PAGE_SIZE in to account.
The following patch attempts to catch the bogus values and round it up to
at least 0-order.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
In the kernels later than 2.6.19 there is a regression that makes swsusp
fail if the resume device is not explicitly specified.
It can be fixed by adding an additional parameter to
mm/swapfile.c:swap_type_of() allowing us to pass the (struct block_device
*) corresponding to the first available swap back to the caller.
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Fix a rather obvious buglet. Noticed while instrumenting the VM using
/proc/vmstat.
Cc: Christoph Lameter <clameter@engr.sgi.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Recent cleanup of slab.h broke SLOB allocator: the routine kmem_cache_init
has now the __init attribute for both slab.c and slob.c. This routine
cannot be removed after init in the case of slob.c -- it serves as a timer
callback.
Provide a separate timer callback routine, call it once from kmem_cache_init,
keep the __init attribute on the latter.
Signed-off-by: Dimitri Gorokhovik <dimitri.gorokhovik@free.fr>
Cc: Christoph Lameter <clameter@engr.sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
constrained_alloc(), which is called to detect where oom is from, checks
passed zone_list(). If zone_list doesn't include all nodes, it thinks oom
is from mempolicy.
But there is memory-less-node. memory-less-node's zones are never included
in zonelist[].
contstrained_alloc() should get memory_less_node into count. Otherwise, it
always thinks 'oom is from mempolicy'. This means that current process
dies at any time. This patch fix it.
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Paul Jackson <pj@sgi.com>
Cc: Christoph Lameter <clameter@engr.sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The VM layer (on the face of it, fairly reasonably) expected that when
it does a ->writepage() call to the filesystem, it would write out the
full page at that point in time. Especially since it had earlier marked
the whole page dirty with "set_page_dirty()".
But that isn't actually the case: ->writepage() does not actually write
a page, it writes the parts of the page that have been explicitly marked
dirty before, *and* that had not got written out for other reasons since
the last time we told it they were dirty.
That last caveat is the important one.
Which _most_ of the time ends up being the whole page (since we had
called "set_page_dirty()" on the page earlier), but if the filesystem
had done any dirty flushing of its own (for example, to honor some
internal write ordering guarantees), it might end up doing only a
partial page IO (or none at all) when ->writepage() is actually called.
That is the correct thing in general (since we actually often _want_
only the known-dirty parts of the page to be written out), but the
shared dirty page handling had implicitly forgotten about these details,
and had a number of cases where it was doing just the "->writepage()"
part, without telling the low-level filesystem that the whole page might
have been re-dirtied as part of being mapped writably into user space.
Since most of the time the FS did actually write out the full page, we
didn't notice this for a loong time, and this needed some really odd
patterns to trigger. But it caused occasional corruption with rtorrent
and with the Debian "apt" database, because both use shared mmaps to
update the end result.
This fixes it. Finally. After way too much hair-pulling.
Acked-by: Nick Piggin <nickpiggin@yahoo.com.au>
Acked-by: Martin J. Bligh <mbligh@google.com>
Acked-by: Martin Michlmayr <tbm@cyrius.com>
Acked-by: Martin Johansson <martin@fatbob.nu>
Acked-by: Ingo Molnar <mingo@elte.hu>
Acked-by: Andrei Popa <andrei.popa@i-neo.ro>
Cc: High Dickins <hugh@veritas.com>
Cc: Andrew Morton <akpm@osdl.org>,
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Segher Boessenkool <segher@kernel.crashing.org>
Cc: David Miller <davem@davemloft.net>
Cc: Arjan van de Ven <arjan@infradead.org>
Cc: Gordon Farquharson <gordonfarquharson@gmail.com>
Cc: Guillaume Chazarain <guichaz@yahoo.fr>
Cc: Theodore Tso <tytso@mit.edu>
Cc: Kenneth Cheng <kenneth.w.chen@intel.com>
Cc: Tobias Diedrich <ranma@tdiedrich.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Make cancel_dirty_page() act more like all the other dirty and writeback
accounting functions: test for "mapping" being NULL, and do the
NR_FILE_DIRY accounting purely based on mapping_cap_account_dirty()).
Also, add it to the exports, so that modular filesystems can use it.
Acked-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Add more debugging in the rmap code in an attempt to locate to source of
the occasional "mapcount went negative" assertions.
Cc: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The version of mm/vmscan.c in Linus' current tree has swapped parameters in
the shrink_all_zones declaration and call, used by the various
suspend-to-disk implementations. This doesn't seem to have any great
adverse effect, but it's clearly wrong.
Signed-off-by: Nigel Cunningham <nigel@suspend2.net>
Cc: Pavel Machek <pavel@ucw.cz>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The declaration of kmem_ptr_validate in slab.h does not match the
one in slab.c. Remove the fastcall attribute (this is the only use in
slab.c).
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Ran into BUG() while doing madvise(REMOVE) testing. If we are punching a
hole into shared memory segment using madvise(REMOVE) and the entire hole
is below the indirect blocks, we hit following assert.
BUG_ON(limit <= SHMEM_NR_DIRECT);
Signed-off-by: Badari Pulavarty <pbadari@us.ibm.com>
Cc: Hugh Dickins <hugh@veritas.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Only (un)account for IO and page-dirtying for devices which have real backing
store (ie: not tmpfs or ramdisks).
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
truncate presently invalidates the dirty page's buffer_heads then shoots down
the page. But try_to_free_buffers() will now bale out because the page is
dirty.
Net effect: the LRU gets filled with dirty pages which have invalidated
buffer_heads attached. They have no ->mapping and hence cannot be cleaned.
The machine leaks memory at an enormous rate.
Fix this by cleaning the page before running try_to_free_buffers(), so
try_to_free_buffers() can do its work.
Also, remember to do dirty-page-acoounting in cancel_dirty_page() so the
machine won't wedge up trying to write non-existent dirty pages.
Probably still wrong, but now less so.
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
They were horribly easy to mis-use because of their tempting naming, and
they also did way more than any users of them generally wanted them to
do.
A dirty page can become clean under two circumstances:
(a) when we write it out. We have "clear_page_dirty_for_io()" for
this, and that function remains unchanged.
In the "for IO" case it is not sufficient to just clear the dirty
bit, you also have to mark the page as being under writeback etc.
(b) when we actually remove a page due to it becoming inaccessible to
users, notably because it was truncate()'d away or the file (or
metadata) no longer exists, and we thus want to cancel any
outstanding dirty state.
For the (b) case, we now introduce "cancel_dirty_page()", which only
touches the page state itself, and verifies that the page is not mapped
(since cancelling writes on a mapped page would be actively wrong as it
is still accessible to users).
Some filesystems need to be fixed up for this: CIFS, FUSE, JFS,
ReiserFS, XFS all use the old confusing functions, and will be fixed
separately in subsequent commits (with some of them just removing the
offending logic, and others using clear_page_dirty_for_io()).
This was confirmed by Martin Michlmayr to fix the apt database
corruption on ARM.
Cc: Martin Michlmayr <tbm@cyrius.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Hugh Dickins <hugh@veritas.com>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: Arjan van de Ven <arjan@infradead.org>
Cc: Andrei Popa <andrei.popa@i-neo.ro>
Cc: Andrew Morton <akpm@osdl.org>
Cc: Dave Kleikamp <shaggy@linux.vnet.ibm.com>
Cc: Gordon Farquharson <gordonfarquharson@gmail.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Trond Myklebust <trond.myklebust@fys.uio.no>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Hugh Dickins correctly points out that mincore() is actually _supposed_
to fail on an unmapped hole in the user address space, rather than
return valid ("empty") information about the hole. This just simplifies
the problem further (I had been misled by our previous confusing and
complicated way of doing mincore()).
Also, in the unlikely situation that we can't allocate a temporary
kernel buffer, we should actually return EAGAIN, not ENOMEM, to keep the
"unmapped hole" and "allocation failure" error cases separate.
Finally, add a comment about our stupid historical lack of support for
anonymous mappings. I'll fix that if somebody reminds me after 2.6.20
is out.
Acked-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Doug Chapman noticed that mincore() will doa "copy_to_user()" of the
result while holding the mmap semaphore for reading, which is a big
no-no. While a recursive read-lock on a semaphore in the case of a page
fault happens to work, we don't actually allow them due to deadlock
schenarios with writers due to fairness issues.
Doug and Marcel sent in a patch to fix it, but I decided to just rewrite
the mess instead - not just fixing the locking problem, but making the
code smaller and (imho) much easier to understand.
Cc: Doug Chapman <dchapman@redhat.com>
Cc: Marcel Holtmann <holtmann@redhat.com>
Cc: Hugh Dickins <hugh@veritas.com>
Cc: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
To allow a more effective copy_user_highpage() on certain architectures,
a vma argument is added to the function and cow_user_page() allowing
the implementation of these functions to check for the VM_EXEC bit.
The main part of this patch was originally written by Ralf Baechle;
Atushi Nemoto did the the debugging.
Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
When some objects are allocated by one CPU but freed by another CPU we can
consume lot of cycles doing divides in obj_to_index().
(Typical load on a dual processor machine where network interrupts are
handled by one particular CPU (allocating skbufs), and the other CPU is
running the application (consuming and freeing skbufs))
Here on one production server (dual-core AMD Opteron 285), I noticed this
divide took 1.20 % of CPU_CLK_UNHALTED events in kernel. But Opteron are
quite modern cpus and the divide is much more expensive on oldest
architectures :
On a 200 MHz sparcv9 machine, the division takes 64 cycles instead of 1
cycle for a multiply.
Doing some math, we can use a reciprocal multiplication instead of a divide.
If we want to compute V = (A / B) (A and B being u32 quantities)
we can instead use :
V = ((u64)A * RECIPROCAL(B)) >> 32 ;
where RECIPROCAL(B) is precalculated to ((1LL << 32) + (B - 1)) / B
Note :
I wrote pure C code for clarity. gcc output for i386 is not optimal but
acceptable :
mull 0x14(%ebx)
mov %edx,%eax // part of the >> 32
xor %edx,%edx // useless
mov %eax,(%esp) // could be avoided
mov %edx,0x4(%esp) // useless
mov (%esp),%ebx
[akpm@osdl.org: small cleanups]
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Cc: Christoph Lameter <clameter@sgi.com>
Cc: David Miller <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Elaborate the API for calling cpuset_zone_allowed(), so that users have to
explicitly choose between the two variants:
cpuset_zone_allowed_hardwall()
cpuset_zone_allowed_softwall()
Until now, whether or not you got the hardwall flavor depended solely on
whether or not you or'd in the __GFP_HARDWALL gfp flag to the gfp_mask
argument.
If you didn't specify __GFP_HARDWALL, you implicitly got the softwall
version.
Unfortunately, this meant that users would end up with the softwall version
without thinking about it. Since only the softwall version might sleep,
this led to bugs with possible sleeping in interrupt context on more than
one occassion.
The hardwall version requires that the current tasks mems_allowed allows
the node of the specified zone (or that you're in interrupt or that
__GFP_THISNODE is set or that you're on a one cpuset system.)
The softwall version, depending on the gfp_mask, might allow a node if it
was allowed in the nearest enclusing cpuset marked mem_exclusive (which
requires taking the cpuset lock 'callback_mutex' to evaluate.)
This patch removes the cpuset_zone_allowed() call, and forces the caller to
explicitly choose between the hardwall and the softwall case.
If the caller wants the gfp_mask to determine this choice, they should (1)
be sure they can sleep or that __GFP_HARDWALL is set, and (2) invoke the
cpuset_zone_allowed_softwall() routine.
This adds another 100 or 200 bytes to the kernel text space, due to the few
lines of nearly duplicate code at the top of both cpuset_zone_allowed_*
routines. It should save a few instructions executed for the calls that
turned into calls of cpuset_zone_allowed_hardwall, thanks to not having to
set (before the call) then check (within the call) the __GFP_HARDWALL flag.
For the most critical call, from get_page_from_freelist(), the same
instructions are executed as before -- the old cpuset_zone_allowed()
routine it used to call is the same code as the
cpuset_zone_allowed_softwall() routine that it calls now.
Not a perfect win, but seems worth it, to reduce this chance of hitting a
sleeping with irq off complaint again.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
More cleanups for slab.h
1. Remove tabs from weird locations as suggested by Pekka
2. Drop the check for NUMA and SLAB_DEBUG from the fallback section
as suggested by Pekka.
3. Uses static inline for the fallback defs as also suggested by Pekka.
4. Make kmem_ptr_valid take a const * argument.
5. Separate the NUMA fallback definitions from the kmalloc_track fallback
definitions.
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Cc: Pekka Enberg <penberg@cs.helsinki.fi>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This is a response to an earlier discussion on linux-mm about splitting
slab.h components per allocator. Patch is against 2.6.19-git11. See
http://marc.theaimsgroup.com/?l=linux-mm&m=116469577431008&w=2
This patch cleans up the slab header definitions. We define the common
functions of slob and slab in slab.h and put the extra definitions needed
for slab's kmalloc implementations in <linux/slab_def.h>. In order to get
a greater set of common functions we add several empty functions to slob.c
and also rename slob's kmalloc to __kmalloc.
Slob does not need any special definitions since we introduce a fallback
case. If there is no need for a slab implementation to provide its own
kmalloc mess^H^H^Hacros then we simply fall back to __kmalloc functions.
That is sufficient for SLOB.
Sort the function in slab.h according to their functionality. First the
functions operating on struct kmem_cache * then the kmalloc related
functions followed by special debug and fallback definitions.
Also redo a lot of comments.
Signed-off-by: Christoph Lameter <clameter@sgi.com>?
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Fallback_alloc() does not do the check for GFP_WAIT as done in
cache_grow(). Thus interrupts are disabled when we call kmem_getpages()
which results in the failure.
Duplicate the handling of GFP_WAIT in cache_grow().
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Cc: Jay Cliburn <jacliburn@bellsouth.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch introduces users of the round_jiffies() function in the slab code.
The slab code has a few "run every second" timers for background work; these
are obviously not timing critical as long as they happen roughly at the right
frequency.
Signed-off-by: Arjan van de Ven <arjan@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The only time it is safe to call aio_complete() is when the ->ki_retry
function returns -EIOCBQUEUED to the AIO core. direct_io_worker() has
historically done this by relying on its caller to translate positive return
codes into -EIOCBQUEUED for the aio case. It did this by trying to keep
conditionals in sync. direct_io_worker() knew when finished_one_bio() was
going to call aio_complete(). It would reverse the test and wait and free the
dio in the cases it thought that finished_one_bio() wasn't going to.
Not surprisingly, it ended up getting it wrong. 'ret' could be a negative
errno from the submission path but it failed to communicate this to
finished_one_bio(). direct_io_worker() would return < 0, it's callers
wouldn't raise -EIOCBQUEUED, and aio_complete() would be called. In the
future finished_one_bio()'s tests wouldn't reflect this and aio_complete()
would be called for a second time which can manifest as an oops.
The previous cleanups have whittled the sync and async completion paths down
to the point where we can collapse them and clearly reassert the invariant
that we must only call aio_complete() after returning -EIOCBQUEUED.
direct_io_worker() will only return -EIOCBQUEUED when it is not the last to
drop the dio refcount and the aio bio completion path will only call
aio_complete() when it is the last to drop the dio refcount.
direct_io_worker() can ensure that it is the last to drop the reference count
by waiting for bios to drain. It does this for sync ops, of course, and for
partial dio writes that must fall back to buffered and for aio ops that saw
errors during submission.
This means that operations that end up waiting, even if they were issued as
aio ops, will not call aio_complete() from dio. Instead we return the return
code of the operation and let the aio core call aio_complete(). This is
purposely done to fix a bug where AIO DIO file extensions would call
aio_complete() before their callers have a chance to update i_size.
Now that direct_io_worker() is explicitly returning -EIOCBQUEUED its callers
no longer have to translate for it. XFS needs to be careful not to free
resources that will be used during AIO completion if -EIOCBQUEUED is returned.
We maintain the previous behaviour of trying to write fs metadata for O_SYNC
aio+dio writes.
Signed-off-by: Zach Brown <zach.brown@oracle.com>
Cc: Badari Pulavarty <pbadari@us.ibm.com>
Cc: Suparna Bhattacharya <suparna@in.ibm.com>
Acked-by: Jeff Moyer <jmoyer@redhat.com>
Cc: <xfs-masters@oss.sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Account for the number of byte writes which this process caused to not happen
after all.
Cc: Jay Lan <jlan@sgi.com>
Cc: Shailabh Nagar <nagar@watson.ibm.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Chris Sturtivant <csturtiv@sgi.com>
Cc: Tony Ernst <tee@sgi.com>
Cc: Guillaume Thouvenin <guillaume.thouvenin@bull.net>
Cc: David Wright <daw@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Accounting writes is fairly simple: whenever a process flips a page from clean
to dirty, we accuse it of having caused a write to underlying storage of
PAGE_CACHE_SIZE bytes.
This may overestimate the amount of writing: the page-dirtying may cause only
one buffer_head's worth of writeout. Fixing that is possible, but probably a
bit messy and isn't obviously important.
Cc: Jay Lan <jlan@sgi.com>
Cc: Shailabh Nagar <nagar@watson.ibm.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Chris Sturtivant <csturtiv@sgi.com>
Cc: Tony Ernst <tee@sgi.com>
Cc: Guillaume Thouvenin <guillaume.thouvenin@bull.net>
Cc: David Wright <daw@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Save a tabstop in __set_page_dirty_nobuffers() and __set_page_dirty_buffers()
and a few other places. No functional changes.
Cc: Jay Lan <jlan@sgi.com>
Cc: Shailabh Nagar <nagar@watson.ibm.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Chris Sturtivant <csturtiv@sgi.com>
Cc: Tony Ernst <tee@sgi.com>
Cc: Guillaume Thouvenin <guillaume.thouvenin@bull.net>
Cc: David Wright <daw@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Ramiro Voicu hits the BUG_ON(!pte_none(*pte)) in zeromap_pte_range: kernel
bugzilla 7645. Right: read_zero_pagealigned uses down_read of mmap_sem,
but another thread's racing read of /dev/zero, or a normal fault, can
easily set that pte again, in between zap_page_range and zeromap_page_range
getting there. It's been wrong ever since 2.4.3.
The simple fix is to use down_write instead, but that would serialize reads
of /dev/zero more than at present: perhaps some app would be badly
affected. So instead let zeromap_page_range return the error instead of
BUG_ON, and read_zero_pagealigned break to the slower clear_user loop in
that case - there's no need to optimize for it.
Use -EEXIST for when a pte is found: BUG_ON in mmap_zero (the other user of
zeromap_page_range), though it really isn't interesting there. And since
mmap_zero wants -EAGAIN for out-of-memory, the zeromaps better return that
than -ENOMEM.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Cc: Ramiro Voicu: <Ramiro.Voicu@cern.ch>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Assign defaults most likely to please a new user:
1) generate some logging output
(verbose=2)
2) avoid injecting failures likely to lock up UI
(ignore_gfp_wait=1, ignore_gfp_highmem=1)
Signed-off-by: Don Mullis <dwm@meer.net>
Cc: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch provides fault-injection capability for alloc_pages()
Boot option:
fail_page_alloc=<interval>,<probability>,<space>,<times>
<interval> -- specifies the interval of failures.
<probability> -- specifies how often it should fail in percent.
<space> -- specifies the size of free space where memory can be
allocated safely in pages.
<times> -- specifies how many times failures may happen at most.
Debugfs:
/debug/fail_page_alloc/interval
/debug/fail_page_alloc/probability
/debug/fail_page_alloc/specifies
/debug/fail_page_alloc/times
/debug/fail_page_alloc/ignore-gfp-highmem
/debug/fail_page_alloc/ignore-gfp-wait
Example:
fail_page_alloc=10,100,0,-1
The page allocation (alloc_pages(), ...) fails once per 10 times.
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch provides fault-injection capability for kmalloc.
Boot option:
failslab=<interval>,<probability>,<space>,<times>
<interval> -- specifies the interval of failures.
<probability> -- specifies how often it should fail in percent.
<space> -- specifies the size of free space where memory can be
allocated safely in bytes.
<times> -- specifies how many times failures may happen at most.
Debugfs:
/debug/failslab/interval
/debug/failslab/probability
/debug/failslab/specifies
/debug/failslab/times
/debug/failslab/ignore-gfp-highmem
/debug/failslab/ignore-gfp-wait
Example:
failslab=10,100,0,-1
slab allocation (kmalloc(), kmem_cache_alloc(),..) fails once per 10 times.
Cc: Pekka Enberg <penberg@cs.helsinki.fi>
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This facility provides three entry points:
ilog2() Log base 2 of unsigned long
ilog2_u32() Log base 2 of u32
ilog2_u64() Log base 2 of u64
These facilities can either be used inside functions on dynamic data:
int do_something(long q)
{
...;
y = ilog2(x)
...;
}
Or can be used to statically initialise global variables with constant values:
unsigned n = ilog2(27);
When performing static initialisation, the compiler will report "error:
initializer element is not constant" if asked to take a log of zero or of
something not reducible to a constant. They treat negative numbers as
unsigned.
When not dealing with a constant, they fall back to using fls() which permits
them to use arch-specific log calculation instructions - such as BSR on
x86/x86_64 or SCAN on FRV - if available.
[akpm@osdl.org: MMC fix]
Signed-off-by: David Howells <dhowells@redhat.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: David Howells <dhowells@redhat.com>
Cc: Wojtek Kaniewski <wojtekka@toxygen.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Change all the uses of f_{dentry,vfsmnt} to f_path.{dentry,mnt} in linux/mm/.
Signed-off-by: Josef "Jeff" Sipek <jsipek@cs.sunysb.edu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
fallback_alloc() could end up calling cpuset_zone_allowed() with interrupts
disabled (by code in kmem_cache_alloc_node()), but without __GFP_HARDWALL
set, leading to a possible call of a sleeping function with interrupts
disabled.
This results in the BUG report:
BUG: sleeping function called from invalid context at kernel/cpuset.c:1520
in_atomic():0, irqs_disabled():1
Thanks to Paul Menage for catching this one.
Signed-off-by: Paul Jackson <pj@sgi.com>
Cc: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
- move some file_operations structs into the .rodata section
- move static strings from policy_types[] array into the .rodata section
- fix generic seq_operations usages, so that those structs may be defined
as "const" as well
[akpm@osdl.org: couple of fixes]
Signed-off-by: Helge Deller <deller@gmx.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
In time for 2.6.20, we can get rid of this junk.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
There was lots of #ifdef noise in the kernel due to hotcpu_notifier(fn,
prio) not correctly marking 'fn' as used in the !HOTPLUG_CPU case, and thus
generating compiler warnings of unused symbols, hence forcing people to add
#ifdefs.
the compiler can skip truly unused functions just fine:
text data bss dec hex filename
1624412 728710 3674856 6027978 5bfaca vmlinux.before
1624412 728710 3674856 6027978 5bfaca vmlinux.after
[akpm@osdl.org: topology.c fix]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
It has no users and it's doubtful that we'll need it again.
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Use put_pages_list() instead of opencoding it.
Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Move process freezing functions from include/linux/sched.h to freezer.h, so
that modifications to the freezer or the kernel configuration don't require
recompiling just about everything.
[akpm@osdl.org: fix ueagle driver]
Signed-off-by: Nigel Cunningham <nigel@suspend2.net>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Currently swsusp saves the contents of highmem pages by copying them to the
normal zone which is quite inefficient (eg. it requires two normal pages
to be used for saving one highmem page). This may be improved by using
highmem for saving the contents of saveable highmem pages.
Namely, during the suspend phase of the suspend-resume cycle we try to
allocate as many free highmem pages as there are saveable highmem pages.
If there are not enough highmem image pages to store the contents of all of
the saveable highmem pages, some of them will be stored in the "normal"
memory. Next, we allocate as many free "normal" pages as needed to store
the (remaining) image data. We use a memory bitmap to mark the allocated
free pages (ie. highmem as well as "normal" image pages).
Now, we use another memory bitmap to mark all of the saveable pages
(highmem as well as "normal") and the contents of the saveable pages are
copied into the image pages. Then, the second bitmap is used to save the
pfns corresponding to the saveable pages and the first one is used to save
their data.
During the resume phase the pfns of the pages that were saveable during the
suspend are loaded from the image and used to mark the "unsafe" page
frames. Next, we try to allocate as many free highmem page frames as to
load all of the image data that had been in the highmem before the suspend
and we allocate so many free "normal" page frames that the total number of
allocated free pages (highmem and "normal") is equal to the size of the
image. While doing this we have to make sure that there will be some extra
free "normal" and "safe" page frames for two lists of PBEs constructed
later.
Now, the image data are loaded, if possible, into their "original" page
frames. The image data that cannot be written into their "original" page
frames are loaded into "safe" page frames and their "original" kernel
virtual addresses, as well as the addresses of the "safe" pages containing
their copies, are stored in one of two lists of PBEs.
One list of PBEs is for the copies of "normal" suspend pages (ie. "normal"
pages that were saveable during the suspend) and it is used in the same way
as previously (ie. by the architecture-dependent parts of swsusp). The
other list of PBEs is for the copies of highmem suspend pages. The pages
in this list are restored (in a reversible way) right before the
arch-dependent code is called.
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Cc: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Make swsusp use block device offsets instead of swap offsets to identify swap
locations and make it use the same code paths for writing as well as for
reading data.
This allows us to use the same code for handling swap files and swap
partitions and to simplify the code, eg. by dropping rw_swap_page_sync().
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Cc: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The Linux kernel handles swap files almost in the same way as it handles swap
partitions and there are only two differences between these two types of swap
areas:
(1) swap files need not be contiguous,
(2) the header of a swap file is not in the first block of the partition
that holds it. From the swsusp's point of view (1) is not a problem,
because it is already taken care of by the swap-handling code, but (2) has
to be taken into consideration.
In principle the location of a swap file's header may be determined with the
help of appropriate filesystem driver. Unfortunately, however, it requires
the filesystem holding the swap file to be mounted, and if this filesystem is
journaled, it cannot be mounted during a resume from disk. For this reason we
need some other means by which swap areas can be identified.
For example, to identify a swap area we can use the partition that holds the
area and the offset from the beginning of this partition at which the swap
header is located.
The following patch allows swsusp to identify swap areas this way. It changes
swap_type_of() so that it takes an additional argument representing an offset
of the swap header within the partition represented by its first argument.
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Make radix tree lookups safe to be performed without locks. Readers are
protected against nodes being deleted by using RCU based freeing. Readers
are protected against new node insertion by using memory barriers to ensure
the node itself will be properly written before it is visible in the radix
tree.
Each radix tree node keeps a record of their height (above leaf nodes).
This height does not change after insertion -- when the radix tree is
extended, higher nodes are only inserted in the top. So a lookup can take
the pointer to what is *now* the root node, and traverse down it even if
the tree is concurrently extended and this node becomes a subtree of a new
root.
"Direct" pointers (tree height of 0, where root->rnode points directly to
the data item) are handled by using the low bit of the pointer to signal
whether rnode is a direct pointer or a pointer to a radix tree node.
When a reader wants to traverse the next branch, they will take a copy of
the pointer. This pointer will be either NULL (and the branch is empty) or
non-NULL (and will point to a valid node).
[akpm@osdl.org: cleanups]
[Lee.Schermerhorn@hp.com: bugfixes, comments, simplifications]
[clameter@sgi.com: build fix]
Signed-off-by: Nick Piggin <npiggin@suse.de>
Cc: "Paul E. McKenney" <paulmck@us.ibm.com>
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Cc: Christoph Lameter <clameter@engr.sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Currently we we use the lru head link of the second page of a compound page
to hold its destructor. This was ok when it was purely an internal
implmentation detail. However, hugetlbfs overrides this destructor
violating the layering. Abstract this out as explicit calls, also
introduce a type for the callback function allowing them to be type
checked. For each callback we pre-declare the function, causing a type
error on definition rather than on use elsewhere.
[akpm@osdl.org: cleanups]
Signed-off-by: Andy Whitcroft <apw@shadowen.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Currently we simply attempt to allocate from all allowed nodes using
GFP_THISNODE. However, GFP_THISNODE does not do reclaim (it wont do any at
all if the recent GFP_THISNODE patch is accepted). If we truly run out of
memory in the whole system then fallback_alloc may return NULL although
memory may still be available if we would perform more thorough reclaim.
This patch changes fallback_alloc() so that we first only inspect all the
per node queues for available slabs. If we find any then we allocate from
those. This avoids slab fragmentation by first getting rid of all partial
allocated slabs on every node before allocating new memory.
If we cannot satisfy the allocation from any per node queue then we extend
a slab. We now call into the page allocator without specifying
GFP_THISNODE. The page allocator will then implement its own fallback (in
the given cpuset context), perform necessary reclaim (again considering not
a single node but the whole set of allowed nodes) and then return pages for
a new slab.
We identify from which node the pages were allocated and then insert the
pages into the corresponding per node structure. In order to do so we need
to modify cache_grow() to take a parameter that specifies the new slab.
kmem_getpages() can no longer set the GFP_THISNODE flag since we need to be
able to use kmem_getpage to allocate from an arbitrary node. GFP_THISNODE
needs to be specified when calling cache_grow().
One key advantage is that the decision from which node to allocate new
memory is removed from slab fallback processing. The patch allows to go
back to use of the page allocators fallback/reclaim logic.
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The intent of GFP_THISNODE is to make sure that an allocation occurs on a
particular node. If this is not possible then NULL needs to be returned so
that the caller can choose what to do next on its own (the slab allocator
depends on that).
However, GFP_THISNODE currently triggers reclaim before returning a failure
(GFP_THISNODE means GFP_NORETRY is set). If we have over allocated a node
then we will currently do some reclaim before returning NULL. The caller
may want memory from other nodes before reclaim should be triggered. (If
the caller wants reclaim then he can directly use __GFP_THISNODE instead).
There is no flag to avoid reclaim in the page allocator and adding yet
another GFP_xx flag would be difficult given that we are out of available
flags.
So just compare and see if all bits for GFP_THISNODE (__GFP_THISNODE,
__GFP_NORETRY and __GFP_NOWARN) are set. If so then we return NULL before
waking up kswapd.
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This addresses two issues:
1. Kmalloc_node() may intermittently return NULL if we are allocating
from the current node and are unable to obtain memory for the current
node from the page allocator. This is because we call ___cache_alloc()
if nodeid == numa_node_id() and ____cache_alloc is not able to fallback
to other nodes.
This was introduced in the 2.6.19 development cycle. <= 2.6.18 in
that case does not do a restricted allocation and blindly trusts the
page allocator to have given us memory from the indicated node. It
inserts the page regardless of the node it came from into the queues for
the current node.
2. If kmalloc_node() is used on a node that has not been bootstrapped
yet then we may try to pass an invalid node number to
____cache_alloc_node() triggering a BUG().
Change the function to call fallback_alloc() instead. Only call
fallback_alloc() if we are allowed to fallback at all. The need to
handle a node not bootstrapped yet also first surfaced in the 2.6.19
cycle.
Update the comments since they were still describing the old kmalloc_node
from 2.6.12.
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
SLAB_DMA is an alias of GFP_DMA. This is the last one so we
remove the leftover comment too.
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
SLAB_KERNEL is an alias of GFP_KERNEL.
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
SLAB_LEVEL_MASK is only used internally to the slab and is
and alias of GFP_LEVEL_MASK.
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
It is only used internally in the slab.
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
David Binderman and his Intel C compiler rightly observe that
install_file_pte no longer has any use for its pte_val.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Cc: d binderman <dcb314@hotmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
These patches introduced new switch statements which are indented contrary
to the concensus in mm/*.c. Fix them up to match that concensus.
[PATCH] node local per-cpu-pages
[PATCH] ZVC: Scale thresholds depending on the size of the system
commit e7c8d5c995
commit df9ecaba3f
Signed-off-by: Andy Whitcroft <apw@shadowen.org>
Cc: Christoph Lameter <clameter@engr.sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The fsfuzzer found this; with a corrupt small swapfile that claims to have
many pages:
[root]# file swap.741.img
swap.741.img: Linux/i386 swap file (new style) 1 (4K pages) size 1040191487 pages
[root]# ls -l swap.741.img
-rw-r--r-- 1 root root 16777216 Nov 22 05:18 swap.741.img
sys_swapon() will try to vmalloc all those pages, and -then- check to see if
the file is actually that large:
if (!(p->swap_map = vmalloc(maxpages * sizeof(short)))) {
<snip>
if (swapfilesize && maxpages > swapfilesize) {
printk(KERN_WARNING
"Swap area shorter than signature indicates\n");
It seems to me that it would make more sense to move this test up before
the vmalloc, with the other checks, to avoid the OOM-killer in this
situation...
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Cc: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
NUMA node ids are passed as either int or unsigned int almost exclusivly
page_to_nid and zone_to_nid both return unsigned long. This is a throw
back to when page_to_nid was a #define and was thus exposing the real type
of the page flags field.
In addition to fixing up the definitions of page_to_nid and zone_to_nid I
audited the users of these functions identifying the following incorrect
uses:
1) mm/page_alloc.c show_node() -- printk dumping the node id,
2) include/asm-ia64/pgalloc.h pgtable_quicklist_free() -- comparison
against numa_node_id() which returns an int from cpu_to_node(), and
3) mm/mpolicy.c check_pte_range -- used as an index in node_isset which
uses bit_set which in generic code takes an int.
Signed-off-by: Andy Whitcroft <apw@shadowen.org>
Cc: Christoph Lameter <clameter@engr.sgi.com>
Cc: "Luck, Tony" <tony.luck@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
drain_node_pages() currently drains the complete pageset of all pages. If
there are a large number of pages in the queues then we may hold off
interrupts for too long.
Duplicate the method used in free_hot_cold_page. Only drain pcp->batch
pages at one time.
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch makes the needlessly global "global_faults" static.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
When booting a NUMA system with nodes that have no memory (eg by limiting
memory), bootmem_alloc_core tried to find pages in an uninitialized
bootmem_map. This caused a null pointer access. This fix adds a check, so
that NULL is returned. That will enable the caller (bootmem_alloc_nopanic)
to alloc memory on other without a panic.
Signed-off-by: Christian Krafft <krafft@de.ibm.com>
Cc: Christoph Lameter <clameter@engr.sgi.com>
Cc: Andy Whitcroft <apw@shadowen.org>
Cc: Martin Bligh <mbligh@google.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The patch (as824b) makes percpu_free() ignore NULL arguments, as one would
expect for a deallocation routine. (Note that free_percpu is #defined as
percpu_free in include/linux/percpu.h.) A few callers are updated to remove
now-unneeded tests for NULL. A few other callers already seem to assume
that passing a NULL pointer to percpu_free() is okay!
The patch also removes an unnecessary NULL check in percpu_depopulate().
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
We have variants of kmalloc and kmem_cache_alloc that leave leak tracking to
the caller. This is used for subsystem-specific allocators like skb_alloc.
To make skb_alloc node-aware we need similar routines for the node-aware slab
allocator, which this patch adds.
Note that the code is rather ugly, but it mirrors the non-node-aware code 1:1:
[akpm@osdl.org: add module export]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
It would be possible for /proc/swaps to not always print out the header:
swapon /dev/hdc2
swapon /dev/hde2
swapoff /dev/hdc2
At this point /proc/swaps would not have a header.
Signed-off-by: Suleiman Souhlal <suleiman@google.com>
Cc: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
OOM can panic due to the processes stuck in __alloc_pages() doing infinite
rebalance loop while no memory can be reclaimed. OOM killer tries to kill
some processes, but unfortunetaly, rebalance label was moved by someone
below the TIF_MEMDIE check, so buddy allocator doesn't see that process is
OOM-killed and it can simply fail the allocation :/
Observed in reality on RHEL4(2.6.9)+OpenVZ kernel when a user doing some
memory allocation tricks triggered OOM panic.
Signed-off-by: Denis Lunev <den@sw.ru>
Signed-off-by: Kirill Korotaev <dev@openvz.org>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
mm is defined as vma->vm_mm, so use that.
Acked-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
When using numa=fake on non-NUMA hardware there is no benefit to having the
alien caches, and they consume much memory.
Add a kernel boot option to disable them.
Christoph sayeth "This is good to have even on large NUMA. The problem is
that the alien caches grow by the square of the size of the system in terms of
nodes."
Cc: Christoph Lameter <clameter@engr.sgi.com>
Cc: Pekka Enberg <penberg@cs.helsinki.fi>
Cc: Manfred Spraul <manfred@colorfullife.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Here's an attempt towards doing away with lock_cpu_hotplug in the slab
subsystem. This approach also fixes a bug which shows up when cpus are
being offlined/onlined and slab caches are being tuned simultaneously.
http://marc.theaimsgroup.com/?l=linux-kernel&m=116098888100481&w=2
The patch has been stress tested overnight on a 2 socket 4 core AMD box with
repeated cpu online and offline, while dbench and kernbench process are
running, and slab caches being tuned at the same time.
There were no lockdep warnings either. (This test on 2,6.18 as 2.6.19-rc
crashes at __drain_pages
http://marc.theaimsgroup.com/?l=linux-kernel&m=116172164217678&w=2 )
The approach here is to hold cache_chain_mutex from CPU_UP_PREPARE until
CPU_ONLINE (similar in approach as worqueue_mutex) . Slab code sensitive
to cpu_online_map (kmem_cache_create, kmem_cache_destroy, slabinfo_write,
__cache_shrink) is already serialized with cache_chain_mutex. (This patch
lengthens cache_chain_mutex hold time at kmem_cache_destroy to cover this).
This patch also takes the cache_chain_sem at kmem_cache_shrink to protect
sanity of cpu_online_map at __cache_shrink, as viewed by slab.
(kmem_cache_shrink->__cache_shrink->drain_cpu_caches). But, really,
kmem_cache_shrink is used at just one place in the acpi subsystem! Do we
really need to keep kmem_cache_shrink at all?
Another note. Looks like a cpu hotplug event can send CPU_UP_CANCELED to
a registered subsystem even if the subsystem did not receive CPU_UP_PREPARE.
This could be due to a subsystem registered for notification earlier than
the current subsystem crapping out with NOTIFY_BAD. Badness can occur with
in the CPU_UP_CANCELED code path at slab if this happens (The same would
apply for workqueue.c as well). To overcome this, we might have to use either
a) a per subsystem flag and avoid handling of CPU_UP_CANCELED, or
b) Use a special notifier events like LOCK_ACQUIRE/RELEASE as Gautham was
using in his experiments, or
c) Do not send CPU_UP_CANCELED to a subsystem which did not receive
CPU_UP_PREPARE.
I would prefer c).
Signed-off-by: Ravikiran Thirumalai <kiran@scalex86.org>
Signed-off-by: Shai Fultheim <shai@scalex86.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
When CONFIG_SLAB_DEBUG is used in combination with ARCH_SLAB_MINALIGN, some
debug flags should be disabled which depend on BYTES_PER_WORD alignment.
The disabling of these debug flags is not properly handled when
BYTES_PER_WORD < ARCH_SLAB_MEMALIGN < cache_line_size()
This patch fixes that and also adds an alignment check to
cache_alloc_debugcheck_after() when ARCH_SLAB_MINALIGN is used.
Signed-off-by: Kevin Hilman <khilman@mvista.com>
Cc: Pekka Enberg <penberg@cs.helsinki.fi>
Cc: Christoph Lameter <clameter@engr.sgi.com>
Cc: Manfred Spraul <manfred@colorfullife.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Imprecise RSS accounting is an irritating ill effect with pt sharing. After
consulted with several VM experts, I have tried various methods to solve that
problem: (1) iterate through all mm_structs that share the PT and increment
count; (2) keep RSS count in page table structure and then sum them up at
reporting time. None of the above methods yield any satisfactory
implementation.
Since process RSS accounting is pure information only, I propose we don't
count them at all for hugetlb page. rlimit has such field, though there is
absolutely no enforcement on limiting that resource. One other method is to
account all RSS at hugetlb mmap time regardless they are faulted or not. I
opt for the simplicity of no accounting at all.
Hugetlb page are special, they are reserved up front in global reservation
pool and is not reclaimable. From physical memory resource point of view, it
is already consumed regardless whether there are users using them.
If the concern is that RSS can be used to control resource allocation, we
already can specify hugetlb fs size limit and sysadmin can enforce that at
mount time. Combined with the two points mentioned above, I fail to see if
there is anything got affected because of this patch.
Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
Acked-by: Hugh Dickins <hugh@veritas.com>
Cc: Dave McCracken <dmccr@us.ibm.com>
Cc: William Lee Irwin III <wli@holomorphy.com>
Cc: "Luck, Tony" <tony.luck@intel.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: David Gibson <david@gibson.dropbear.id.au>
Cc: Adam Litke <agl@us.ibm.com>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Following up with the work on shared page table done by Dave McCracken. This
set of patch target shared page table for hugetlb memory only.
The shared page table is particular useful in the situation of large number of
independent processes sharing large shared memory segments. In the normal
page case, the amount of memory saved from process' page table is quite
significant. For hugetlb, the saving on page table memory is not the primary
objective (as hugetlb itself already cuts down page table overhead
significantly), instead, the purpose of using shared page table on hugetlb is
to allow faster TLB refill and smaller cache pollution upon TLB miss.
With PT sharing, pte entries are shared among hundreds of processes, the cache
consumption used by all the page table is smaller and in return, application
gets much higher cache hit ratio. One other effect is that cache hit ratio
with hardware page walker hitting on pte in cache will be higher and this
helps to reduce tlb miss latency. These two effects contribute to higher
application performance.
Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
Acked-by: Hugh Dickins <hugh@veritas.com>
Cc: Dave McCracken <dmccr@us.ibm.com>
Cc: William Lee Irwin III <wli@holomorphy.com>
Cc: "Luck, Tony" <tony.luck@intel.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: David Gibson <david@gibson.dropbear.id.au>
Cc: Adam Litke <agl@us.ibm.com>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Add an arch_alloc_page to match arch_free_page.
Signed-off-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The new swap token patches replace the current token traversal algo. The old
algo had a crude timeout parameter that was used to handover the token from
one task to another. This algo, transfers the token to the tasks that are in
need of the token. The urgency for the token is based on the number of times
a task is required to swap-in pages. Accordingly, the priority of a task is
incremented if it has been badly affected due to swap-outs. To ensure that
the token doesnt bounce around rapidly, the token holders are given a priority
boost. The priority of tasks is also decremented, if their rate of swap-in's
keeps reducing. This way, the condition to check whether to pre-empt the swap
token, is a matter of comparing two task's priority fields.
[akpm@osdl.org: cleanups]
Signed-off-by: Ashwin Chaugule <ashwin.chaugule@celunite.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Make sure the contention for the token happens _before_ any read-in and
kicks the swap-token algo only when the VM is under pressure.
Signed-off-by: Ashwin Chaugule <ashwin.chaugule@celunite.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Don't cause all threads in all other thread groups to gain TIF_MEMDIE
otherwise we'll get a thundering herd eating our memory reserve. This may not
be the optimal scheme, but it fits our policy of allowing just one TIF_MEMDIE
in the system at once.
Signed-off-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Clean up the OOM killer messages to be more consistent.
Signed-off-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Abort the kill if any of our threads have OOM_DISABLE set. Having this
test here also prevents any OOM_DISABLE child of the "selected" process
from being killed.
Signed-off-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Optimize the critical zonelist scanning for free pages in the kernel memory
allocator by caching the zones that were found to be full recently, and
skipping them.
Remembers the zones in a zonelist that were short of free memory in the
last second. And it stashes a zone-to-node table in the zonelist struct,
to optimize that conversion (minimize its cache footprint.)
Recent changes:
This differs in a significant way from a similar patch that I
posted a week ago. Now, instead of having a nodemask_t of
recently full nodes, I have a bitmask of recently full zones.
This solves a problem that last weeks patch had, which on
systems with multiple zones per node (such as DMA zone) would
take seeing any of these zones full as meaning that all zones
on that node were full.
Also I changed names - from "zonelist faster" to "zonelist cache",
as that seemed to better convey what we're doing here - caching
some of the key zonelist state (for faster access.)
See below for some performance benchmark results. After all that
discussion with David on why I didn't need them, I went and got
some ;). I wanted to verify that I had not hurt the normal case
of memory allocation noticeably. At least for my one little
microbenchmark, I found (1) the normal case wasn't affected, and
(2) workloads that forced scanning across multiple nodes for
memory improved up to 10% fewer System CPU cycles and lower
elapsed clock time ('sys' and 'real'). Good. See details, below.
I didn't have the logic in get_page_from_freelist() for various
full nodes and zone reclaim failures correct. That should be
fixed up now - notice the new goto labels zonelist_scan,
this_zone_full, and try_next_zone, in get_page_from_freelist().
There are two reasons I persued this alternative, over some earlier
proposals that would have focused on optimizing the fake numa
emulation case by caching the last useful zone:
1) Contrary to what I said before, we (SGI, on large ia64 sn2 systems)
have seen real customer loads where the cost to scan the zonelist
was a problem, due to many nodes being full of memory before
we got to a node we could use. Or at least, I think we have.
This was related to me by another engineer, based on experiences
from some time past. So this is not guaranteed. Most likely, though.
The following approach should help such real numa systems just as
much as it helps fake numa systems, or any combination thereof.
2) The effort to distinguish fake from real numa, using node_distance,
so that we could cache a fake numa node and optimize choosing
it over equivalent distance fake nodes, while continuing to
properly scan all real nodes in distance order, was going to
require a nasty blob of zonelist and node distance munging.
The following approach has no new dependency on node distances or
zone sorting.
See comment in the patch below for a description of what it actually does.
Technical details of note (or controversy):
- See the use of "zlc_active" and "did_zlc_setup" below, to delay
adding any work for this new mechanism until we've looked at the
first zone in zonelist. I figured the odds of the first zone
having the memory we needed were high enough that we should just
look there, first, then get fancy only if we need to keep looking.
- Some odd hackery was needed to add items to struct zonelist, while
not tripping up the custom zonelists built by the mm/mempolicy.c
code for MPOL_BIND. My usual wordy comments below explain this.
Search for "MPOL_BIND".
- Some per-node data in the struct zonelist is now modified frequently,
with no locking. Multiple CPU cores on a node could hit and mangle
this data. The theory is that this is just performance hint data,
and the memory allocator will work just fine despite any such mangling.
The fields at risk are the struct 'zonelist_cache' fields 'fullzones'
(a bitmask) and 'last_full_zap' (unsigned long jiffies). It should
all be self correcting after at most a one second delay.
- This still does a linear scan of the same lengths as before. All
I've optimized is making the scan faster, not algorithmically
shorter. It is now able to scan a compact array of 'unsigned
short' in the case of many full nodes, so one cache line should
cover quite a few nodes, rather than each node hitting another
one or two new and distinct cache lines.
- If both Andi and Nick don't find this too complicated, I will be
(pleasantly) flabbergasted.
- I removed the comment claiming we only use one cachline's worth of
zonelist. We seem, at least in the fake numa case, to have put the
lie to that claim.
- I pay no attention to the various watermarks and such in this performance
hint. A node could be marked full for one watermark, and then skipped
over when searching for a page using a different watermark. I think
that's actually quite ok, as it will tend to slightly increase the
spreading of memory over other nodes, away from a memory stressed node.
===============
Performance - some benchmark results and analysis:
This benchmark runs a memory hog program that uses multiple
threads to touch alot of memory as quickly as it can.
Multiple runs were made, touching 12, 38, 64 or 90 GBytes out of
the total 96 GBytes on the system, and using 1, 19, 37, or 55
threads (on a 56 CPU system.) System, user and real (elapsed)
timings were recorded for each run, shown in units of seconds,
in the table below.
Two kernels were tested - 2.6.18-mm3 and the same kernel with
this zonelist caching patch added. The table also shows the
percentage improvement the zonelist caching sys time is over
(lower than) the stock *-mm kernel.
number 2.6.18-mm3 zonelist-cache delta (< 0 good) percent
GBs N ------------ -------------- ---------------- systime
mem threads sys user real sys user real sys user real better
12 1 153 24 177 151 24 176 -2 0 -1 1%
12 19 99 22 8 99 22 8 0 0 0 0%
12 37 111 25 6 112 25 6 1 0 0 -0%
12 55 115 25 5 110 23 5 -5 -2 0 4%
38 1 502 74 576 497 73 570 -5 -1 -6 0%
38 19 426 78 48 373 76 39 -53 -2 -9 12%
38 37 544 83 36 547 82 36 3 -1 0 -0%
38 55 501 77 23 511 80 24 10 3 1 -1%
64 1 917 125 1042 890 124 1014 -27 -1 -28 2%
64 19 1118 138 119 965 141 103 -153 3 -16 13%
64 37 1202 151 94 1136 150 81 -66 -1 -13 5%
64 55 1118 141 61 1072 140 58 -46 -1 -3 4%
90 1 1342 177 1519 1275 174 1450 -67 -3 -69 4%
90 19 2392 199 192 2116 189 176 -276 -10 -16 11%
90 37 3313 238 175 2972 225 145 -341 -13 -30 10%
90 55 1948 210 104 1843 213 100 -105 3 -4 5%
Notes:
1) This test ran a memory hog program that started a specified number N of
threads, and had each thread allocate and touch 1/N'th of
the total memory to be used in the test run in a single loop,
writing a constant word to memory, one store every 4096 bytes.
Watching this test during some earlier trial runs, I would see
each of these threads sit down on one CPU and stay there, for
the remainder of the pass, a different CPU for each thread.
2) The 'real' column is not comparable to the 'sys' or 'user' columns.
The 'real' column is seconds wall clock time elapsed, from beginning
to end of that test pass. The 'sys' and 'user' columns are total
CPU seconds spent on that test pass. For a 19 thread test run,
for example, the sum of 'sys' and 'user' could be up to 19 times the
number of 'real' elapsed wall clock seconds.
3) Tests were run on a fresh, single-user boot, to minimize the amount
of memory already in use at the start of the test, and to minimize
the amount of background activity that might interfere.
4) Tests were done on a 56 CPU, 28 Node system with 96 GBytes of RAM.
5) Notice that the 'real' time gets large for the single thread runs, even
though the measured 'sys' and 'user' times are modest. I'm not sure what
that means - probably something to do with it being slow for one thread to
be accessing memory along ways away. Perhaps the fake numa system, running
ostensibly the same workload, would not show this substantial degradation
of 'real' time for one thread on many nodes -- lets hope not.
6) The high thread count passes (one thread per CPU - on 55 of 56 CPUs)
ran quite efficiently, as one might expect. Each pair of threads needed
to allocate and touch the memory on the node the two threads shared, a
pleasantly parallizable workload.
7) The intermediate thread count passes, when asking for alot of memory forcing
them to go to a few neighboring nodes, improved the most with this zonelist
caching patch.
Conclusions:
* This zonelist cache patch probably makes little difference one way or the
other for most workloads on real numa hardware, if those workloads avoid
heavy off node allocations.
* For memory intensive workloads requiring substantial off-node allocations
on real numa hardware, this patch improves both kernel and elapsed timings
up to ten per-cent.
* For fake numa systems, I'm optimistic, but will have to leave that up to
Rohit Seth to actually test (once I get him a 2.6.18 backport.)
Signed-off-by: Paul Jackson <pj@sgi.com>
Cc: Rohit Seth <rohitseth@google.com>
Cc: Christoph Lameter <clameter@engr.sgi.com>
Cc: David Rientjes <rientjes@cs.washington.edu>
Cc: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
