Replace kmalloc+memset with kzalloc
Signed-off-by: NYan Burman <burman.yan@gmail.com>
Cc: Roman Zippel <zippel@linux-m68k.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Make the workqueues used by XFS freezeable, so their worker threads don't
submit any I/O after the suspend image has been created.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Acked-by: NPavel Machek <pavel@ucw.cz>
Cc: Nigel Cunningham <nigel@suspend2.net>
Cc: David Chinner <dgc@sgi.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Make it possible to create a workqueue the worker thread of which will be
frozen during suspend, along with other kernel threads.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Acked-by: NPavel Machek <pavel@ucw.cz>
Cc: Nigel Cunningham <nigel@suspend2.net>
Cc: David Chinner <dgc@sgi.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Cleanup write-only variable, suggested by D Binderman.
Signed-off-by: NPavel Machek <pavel@suse.cz>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

The 'testproc' swsusp debug mode thaws tasks twice in a row, which is _very_
confusing.  Fix that.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Acked-by: NPavel Machek <pavel@ucw.cz>
Cc: <stable@kernel.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Linus posted quite nice TRACE_RESUME how-to, and I think it is too nice to
be hidden in archives of mailing list, so I turned it into Documentation
piece.
Signed-off-by: NPavel Machek <pavel@suse.cz>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Move all labels in the swsusp code to the second column, so that they won't
fool diff -p.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Acked-by: NPavel Machek <pavel@ucw.cz>
Cc: Nigel Cunningham <nigel@suspend2.net>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Fix coding style in suspend.c.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Acked-by: NPavel Machek <pavel@ucw.cz>
Cc: Nigel Cunningham <nigel@suspend2.net>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Move the loop from freeze_processes() to a separate function and call it
independently for user space processes and kernel threads so that the order
of freezing tasks is clearly visible.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Cc: Pavel Machek <pavel@ucw.cz>
Cc: Nigel Cunningham <nigel@suspend2.net>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Move the loop from thaw_processes() to a separate function and call it
independently for kernel threads and user space processes so that the order
of thawing tasks is clearly visible.

Drop thaw_kernel_threads() which is never used.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Cc: Pavel Machek <pavel@ucw.cz>
Cc: Nigel Cunningham <nigel@suspend2.net>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

The power management semaphore is only used as mutex, so convert it.

[akpm@osdl.org: fix rotten bug]
Signed-off-by: NStephen Hemminger <shemminger@osdl.org>
Acked-by: NIngo Molnar <mingo@elte.hu>
Acked-by: NPavel Machek <pavel@ucw.cz>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Fix http://bugzilla.kernel.org/show_bug.cgi?id=7534

Fix the freezing of processes so that it won't fail if there is a traced
process the parent of which has been stopped.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Acked-by: NPavel Machek <pavel@ucw.cz>
Cc: maurice barnum <pixi+kbug@burble.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Make swsusp measure and print the time needed to shrink memory during the
suspend.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Cc: Pavel Machek <pavel@ucw.cz>
Cc: Nigel Cunningham <nigel@suspend2.net>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Don't modify the cpus_allowed of the task initiating the suspend.
_cpu_down() already makes sure that the task doing the suspend doesn't run
on dying cpu.
Signed-off-by: NSuresh Siddha <suresh.b.siddha@intel.com>
Cc: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Cc: Pavel Machek <pavel@ucw.cz>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: Nigel Cunningham <nigel@suspend2.net>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Make swsusp support i386 systems with PAE or without PSE.

This is done by creating temporary page tables located in resume-safe page
frames before the suspend image is restored in the same way as x86_64 does
it.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Cc: Andi Kleen <ak@suse.de>
Cc: Dave Jones <davej@redhat.com>
Cc: Nigel Cunningham <ncunningham@linuxmail.org>
Cc: Pavel Machek <pavel@ucw.cz>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Modify process thawing so that we can thaw kernel space without thawing
userspace, and thaw kernelspace first.  This will be useful in later
patches, where I intend to get swsusp thawing kernel threads only before
seeking to free memory.
Signed-off-by: NNigel Cunningham <nigel@suspend2.net>
Cc: Pavel Machek <pavel@ucw.cz>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Minor whitespace and formatting modifications for the freezer.
Signed-off-by: NNigel Cunningham <nigel@suspend2.net>
Acked-by: NPavel Machek <pavel@ucw.cz>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

The freezer currently prints an '=' for every process that is frozen.  This
is pretty pointless, as the equals sign says nothing about which process is
frozen, and makes logs look messier (especially if there were a large
number of processes running).  All we really need to know is that we
started trying to freeze processes and what processes (if any) failed to
freeze, or that we succeeded.
Signed-off-by: NNigel Cunningham <nigel@suspend2.net>
Acked-by: NPavel Machek <pavel@ucw.cz>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus 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: NNigel Cunningham <nigel@suspend2.net>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: Pavel Machek <pavel@ucw.cz>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

At some point after 2.6.13, in-kernel software suspend got "incomplete" for
the so-called "platform" mode.  pm_ops->prepare() is never called.  A
visible sign of this is the "moon" light on thinkpads not flashing during
suspend.  Fix by readding the pm_ops->prepare call during suspend.
Signed-off-by: NStefan Seyfried <seife@suse.de>
Acked-by: N"Rafael J. Wysocki" <rjw@sisk.pl>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

swsusp uses GFP_ATOMIC, but it can afford to use __GFP_WAIT, which will
permit it to reclaim clean pagecache instead of emitting scary
page-allocation-failure messages.

Cc: Pavel Machek <pavel@ucw.cz>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus 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: NRafael J. Wysocki <rjw@sisk.pl>
Cc: Pavel Machek <pavel@ucw.cz>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

The swsusp userland interface has recently changed for a couple of times, but
the changes have not been documented.  Fix this, and document the
SNAPSHOT_SET_SWAP_AREA ioctl().
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Acked-by: NPavel Machek <pavel@ucw.cz>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

To be able to use swap files as suspend storage from the userland suspend
tools we need an additional ioctl() that will allow us to provide the kernel
with both the swap header's offset and the identification of the resume
partition.

The new ioctl() should be regarded as a replacement for the
SNAPSHOT_SET_SWAP_FILE ioctl() that from now on will be considered as
obsolete, but has to stay for backwards compatibility of the interface.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Acked-by: NPavel Machek <pavel@ucw.cz>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Document the "resume_offset=" command line parameter as well as the way in
which swap files are supported by swsusp.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Cc: Pavel Machek <pavel@ucw.cz>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Add the kernel command line parameter "resume_offset=" allowing us to specify
the offset, in <PAGE_SIZE> units, from the beginning of the partition pointed
to by the "resume=" parameter at which the swap header is located.

This offset can be determined, for example, by an application using the FIBMAP
ioctl to obtain the swap header's block number for given file.

[akpm@osdl.org: we don't know what type sector_t is]
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Cc: Pavel Machek <pavel@ucw.cz>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus 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: NRafael J. Wysocki <rjw@sisk.pl>
Cc: Pavel Machek <pavel@ucw.cz>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Rearrange the code in kernel/power/swap.c so that the next patch is more
readable.

[This patch only moves the existing code.]
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Acked-by: NPavel Machek <pavel@ucw.cz>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus 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: NRafael J. Wysocki <rjw@sisk.pl>
Acked-by: NPavel Machek <pavel@ucw.cz>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Add an ioctl to the userspace swsusp code that enables the usage of the
pmops->prepare, pmops->enter and pmops->finish methods (the in-kernel
suspend knows these as "platform method").  These are needed on many
machines to (among others) speed up resuming by letting the BIOS skip some
steps or let my hp nx5000 recognise the correct ac_adapter state after
resume again.

It also ensures on many machines, that changed hardware (unplugged AC
adapters) gets correctly detected and that kacpid does not run wild after
resume.
Signed-off-by: NStefan Seyfried <seife@suse.de>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: Pavel Machek <pavel@ucw.cz>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Now that we have pci_get_bus_and_slot we can do the job correctly.  Note that
some of these calls intentionally leak a device - this is because the device
in question is always needed from boot to reboot.
Signed-off-by: NAlan Cox <alan@redhat.com>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Signed-off-by: NMariusz Kozlowski <m.kozlowski@tuxland.pl>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

While working on SH kprobes, I noticed that avr32 got the preemption
handling wrong in the no probe case.  The idea is that upon entry of
kprobe_handler() preemption is disabled outright across the life of the
kprobe, only to be re-enabled in post_kprobe_handler().

However, in the event that the probe is never activated, there's never any
chance of hitting the post probe handler, which allows for the current
avr32 implementation to disable preemption indefinitely, as it's currently
missing a re-enable when no probe is activated.
Signed-off-by: NPaul Mundt <lethal@linux-sh.org>
Cc: Haavard Skinnemoen <hskinnemoen@atmel.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

This patch fixes the following compile error with
-Werror-implicit-function-declaration
(without -Werror-implicit-function-declaration it's a link error):

    ...
      CC      arch/frv/kernel/futex.o
    /home/bunk/linux/kernel-2.6/linux-2.6.19-rc6-mm2/arch/frv/kernel/futex.c:
    In function 'futex_atomic_op_inuser':
    /home/bunk/linux/kernel-2.6/linux-2.6.19-rc6-mm2/arch/frv/kernel/futex.c:203:
    error: implicit declaration of function 'pagefault_disable'
    /home/bunk/linux/kernel-2.6/linux-2.6.19-rc6-mm2/arch/frv/kernel/futex.c:226:
    error: implicit declaration of function 'pagefault_enable'
    make[2]: *** [arch/frv/kernel/futex.o] Error 1
    ...
Signed-off-by: NAdrian Bunk <bunk@stusta.de>
Acked-by: NDavid Howells <dhowells@redhat.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Signed-off-by: NEric Sesterhenn <snakebyte@gmx.de>
Signed-off-by: NAlexey Dobriyan <adobriyan@gmail.com>
Acked-By: NDavid Howells <dhowells@redhat.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus 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: NNick Piggin <npiggin@suse.de>
Cc: "Paul E. McKenney" <paulmck@us.ibm.com>
Signed-off-by: NLee Schermerhorn <lee.schermerhorn@hp.com>
Cc: Christoph Lameter <clameter@engr.sgi.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Before:
[acme@newtoy net-2.6.20]$ pahole --cacheline 32 kernel/sched.o mm_struct

/* include2/asm/processor.h:542 */
struct mm_struct {
        struct vm_area_struct *    mmap;                 /*     0     4 */
        struct rb_root             mm_rb;                /*     4     4 */
        struct vm_area_struct *    mmap_cache;           /*     8     4 */
        long unsigned int          (*get_unmapped_area)(); /*    12     4 */
        void                       (*unmap_area)();      /*    16     4 */
        long unsigned int          mmap_base;            /*    20     4 */
        long unsigned int          task_size;            /*    24     4 */
        long unsigned int          cached_hole_size;     /*    28     4 */
        /* ---------- cacheline 1 boundary ---------- */
        long unsigned int          free_area_cache;      /*    32     4 */
        pgd_t *                    pgd;                  /*    36     4 */
        atomic_t                   mm_users;             /*    40     4 */
        atomic_t                   mm_count;             /*    44     4 */
        int                        map_count;            /*    48     4 */
        struct rw_semaphore        mmap_sem;             /*    52    64 */
        spinlock_t                 page_table_lock;      /*   116    40 */
        struct list_head           mmlist;               /*   156     8 */
        mm_counter_t               _file_rss;            /*   164     4 */
        mm_counter_t               _anon_rss;            /*   168     4 */
        long unsigned int          hiwater_rss;          /*   172     4 */
        long unsigned int          hiwater_vm;           /*   176     4 */
        long unsigned int          total_vm;             /*   180     4 */
        long unsigned int          locked_vm;            /*   184     4 */
        long unsigned int          shared_vm;            /*   188     4 */
        /* ---------- cacheline 6 boundary ---------- */
        long unsigned int          exec_vm;              /*   192     4 */
        long unsigned int          stack_vm;             /*   196     4 */
        long unsigned int          reserved_vm;          /*   200     4 */
        long unsigned int          def_flags;            /*   204     4 */
        long unsigned int          nr_ptes;              /*   208     4 */
        long unsigned int          start_code;           /*   212     4 */
        long unsigned int          end_code;             /*   216     4 */
        long unsigned int          start_data;           /*   220     4 */
        /* ---------- cacheline 7 boundary ---------- */
        long unsigned int          end_data;             /*   224     4 */
        long unsigned int          start_brk;            /*   228     4 */
        long unsigned int          brk;                  /*   232     4 */
        long unsigned int          start_stack;          /*   236     4 */
        long unsigned int          arg_start;            /*   240     4 */
        long unsigned int          arg_end;              /*   244     4 */
        long unsigned int          env_start;            /*   248     4 */
        long unsigned int          env_end;              /*   252     4 */
        /* ---------- cacheline 8 boundary ---------- */
        long unsigned int          saved_auxv[44];       /*   256   176 */
        unsigned int               dumpable:2;           /*   432     4 */
        cpumask_t                  cpu_vm_mask;          /*   436     4 */
        mm_context_t               context;              /*   440    68 */
        long unsigned int          swap_token_time;      /*   508     4 */
        /* ---------- cacheline 16 boundary ---------- */
        char                       recent_pagein;        /*   512     1 */

        /* XXX 3 bytes hole, try to pack */

        int                        core_waiters;         /*   516     4 */
        struct completion *        core_startup_done;    /*   520     4 */
        struct completion          core_done;            /*   524    52 */
        rwlock_t                   ioctx_list_lock;      /*   576    36 */
        struct kioctx *            ioctx_list;           /*   612     4 */
}; /* size: 616, sum members: 613, holes: 1, sum holes: 3, cachelines: 20,
      last cacheline: 8 bytes */

After:

[acme@newtoy net-2.6.20]$ pahole --cacheline 32 kernel/sched.o mm_struct
/* include2/asm/processor.h:542 */
struct mm_struct {
        struct vm_area_struct *    mmap;                 /*     0     4 */
        struct rb_root             mm_rb;                /*     4     4 */
        struct vm_area_struct *    mmap_cache;           /*     8     4 */
        long unsigned int          (*get_unmapped_area)(); /*    12     4 */
        void                       (*unmap_area)();      /*    16     4 */
        long unsigned int          mmap_base;            /*    20     4 */
        long unsigned int          task_size;            /*    24     4 */
        long unsigned int          cached_hole_size;     /*    28     4 */
        /* ---------- cacheline 1 boundary ---------- */
        long unsigned int          free_area_cache;      /*    32     4 */
        pgd_t *                    pgd;                  /*    36     4 */
        atomic_t                   mm_users;             /*    40     4 */
        atomic_t                   mm_count;             /*    44     4 */
        int                        map_count;            /*    48     4 */
        struct rw_semaphore        mmap_sem;             /*    52    64 */
        spinlock_t                 page_table_lock;      /*   116    40 */
        struct list_head           mmlist;               /*   156     8 */
        mm_counter_t               _file_rss;            /*   164     4 */
        mm_counter_t               _anon_rss;            /*   168     4 */
        long unsigned int          hiwater_rss;          /*   172     4 */
        long unsigned int          hiwater_vm;           /*   176     4 */
        long unsigned int          total_vm;             /*   180     4 */
        long unsigned int          locked_vm;            /*   184     4 */
        long unsigned int          shared_vm;            /*   188     4 */
        /* ---------- cacheline 6 boundary ---------- */
        long unsigned int          exec_vm;              /*   192     4 */
        long unsigned int          stack_vm;             /*   196     4 */
        long unsigned int          reserved_vm;          /*   200     4 */
        long unsigned int          def_flags;            /*   204     4 */
        long unsigned int          nr_ptes;              /*   208     4 */
        long unsigned int          start_code;           /*   212     4 */
        long unsigned int          end_code;             /*   216     4 */
        long unsigned int          start_data;           /*   220     4 */
        /* ---------- cacheline 7 boundary ---------- */
        long unsigned int          end_data;             /*   224     4 */
        long unsigned int          start_brk;            /*   228     4 */
        long unsigned int          brk;                  /*   232     4 */
        long unsigned int          start_stack;          /*   236     4 */
        long unsigned int          arg_start;            /*   240     4 */
        long unsigned int          arg_end;              /*   244     4 */
        long unsigned int          env_start;            /*   248     4 */
        long unsigned int          env_end;              /*   252     4 */
        /* ---------- cacheline 8 boundary ---------- */
        long unsigned int          saved_auxv[44];       /*   256   176 */
        cpumask_t                  cpu_vm_mask;          /*   432     4 */
        mm_context_t               context;              /*   436    68 */
        long unsigned int          swap_token_time;      /*   504     4 */
        char                       recent_pagein;        /*   508     1 */
        unsigned char              dumpable:2;           /*   509     1 */

        /* XXX 2 bytes hole, try to pack */

        int                        core_waiters;         /*   512     4 */
        struct completion *        core_startup_done;    /*   516     4 */
        struct completion          core_done;            /*   520    52 */
        rwlock_t                   ioctx_list_lock;      /*   572    36 */
        struct kioctx *            ioctx_list;           /*   608     4 */
}; /* size: 612, sum members: 610, holes: 1, sum holes: 2, cachelines: 20,
      last cacheline: 4 bytes */

[acme@newtoy net-2.6.20]$ codiff -V /tmp/sched.o.before kernel/sched.o
/pub/scm/linux/kernel/git/acme/net-2.6.20/kernel/sched.c:
  struct mm_struct |   -4
    dumpable:2;
     from: unsigned int          /*   432(30)    4(2) */
     to:   unsigned char         /*   509(6)     1(2) */
< SNIP other offset changes >
 1 struct changed
[acme@newtoy net-2.6.20]$

I'm not aware of any problem about using 2 byte wide bitfields where
previously a 4 byte wide one was, holler if there is any, I wouldn't be
surprised, bitfields are things from hell.

For the curious, 432(30) means: at offset 432 from the struct start, at
offset 30 in the bitfield (yeah, it comes backwards, hellish, huh?) ditto
for 509(6), while 4(2) and 1(2) means "struct field size(bitfield size)".

Now we have a 2 bytes hole and are using only 4 bytes of the last 32
bytes cacheline, any takers? :-)
Signed-off-by: NArnaldo Carvalho de Melo <acme@mandriva.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus 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: NAndy Whitcroft <apw@shadowen.org>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus 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: NChristoph Lameter <clameter@sgi.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus 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: NChristoph Lameter <clameter@sgi.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>