When we decide to wake up readers, we must first grant them as many read
locks as necessary, and then actually wake up all these readers.  But in
order to know how many read shares to grant, we must first count the
readers at the head of the queue.  This might take a while if there are
many readers, and we want to be protected against a writer stealing the
lock while we're counting.  To that end, we grant the first reader lock
before counting how many more readers are queued.

We also require some adjustments to the wake_type semantics.

RWSEM_WAKE_NO_ACTIVE used to mean that we had found the count to be
RWSEM_WAITING_BIAS, in which case the rwsem was known to be free as
nobody could steal it while we hold the wait_lock.  This doesn't make
sense once we implement fastpath write lock stealing, so we now use
RWSEM_WAKE_ANY in that case.

Similarly, when rwsem_down_write_failed found that a read lock was
active, it would use RWSEM_WAKE_READ_OWNED which signalled that new
readers could be woken without checking first that the rwsem was
available.  We can't do that anymore since the existing readers might
release their read locks, and a writer could steal the lock before we
wake up additional readers.  So, we have to use a new RWSEM_WAKE_READERS
value to indicate we only want to wake readers, but we don't currently
hold any read lock.
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Peter Hurley <peter@hurleysoftware.com>
Acked-by: Davidlohr Bueso <davidlohr.bueso@hp.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

This is mostly for cleanup value:

- We don't need several gotos to handle the case where the first
  waiter is a writer. Two simple tests will do (and generate very
  similar code).

- In the remainder of the function, we know the first waiter is a reader,
  so we don't have to double check that. We can use do..while loops
  to iterate over the readers to wake (generates slightly better code).
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Peter Hurley <peter@hurleysoftware.com>
Acked-by: Davidlohr Bueso <davidlohr.bueso@hp.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

We can skip the initial trylock in rwsem_down_write_failed() if there
are known active lockers already, thus saving one likely-to-fail
cmpxchg.
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Peter Hurley <peter@hurleysoftware.com>
Acked-by: Davidlohr Bueso <davidlohr.bueso@hp.com>
Acked-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

In rwsem_down_write_failed(), if there are active locks after we wake up
(i.e.  the lock got stolen from us), skip taking the wait_lock and go
back to sleep immediately.
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Peter Hurley <peter@hurleysoftware.com>
Acked-by: Davidlohr Bueso <davidlohr.bueso@hp.com>
Acked-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Using rwsem_atomic_update to try stealing the write lock forced us to
undo the adjustment in the failure path.  We can have simpler and faster
code by using cmpxchg instead.
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Peter Hurley <peter@hurleysoftware.com>
Acked-by: Davidlohr Bueso <davidlohr.bueso@hp.com>
Acked-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Some small code simplifications can be achieved by doing more agressive
lock stealing:

- When rwsem_down_write_failed() notices that there are no active locks
  (and thus no thread to wake us if we decided to sleep), it used to wake
  the first queued process. However, stealing the lock is also sufficient
  to deal with this case, so we don't need this check anymore.

- In try_get_writer_sem(), we can steal the lock even when the first waiter
  is a reader. This is correct because the code path that wakes readers is
  protected by the wait_lock. As to the performance effects of this change,
  they are expected to be minimal: readers are still granted the lock
  (rather than having to acquire it themselves) when they reach the front
  of the wait queue, so we have essentially the same behavior as in
  rwsem-spinlock.
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Reviewed-by: Peter Hurley <peter@hurleysoftware.com>
Acked-by: Davidlohr Bueso <davidlohr.bueso@hp.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

When waking writers, we never grant them the lock - instead, they have
to acquire it themselves when they run, and remove themselves from the
wait_list when they succeed.

As a result, we can do a few simplifications in rwsem_down_write_failed():

- We don't need to check for !waiter.task since __rwsem_do_wake() doesn't
  remove writers from the wait_list

- There is no point releaseing the wait_lock before entering the wait loop,
  as we will need to reacquire it immediately. We can change the loop so
  that the lock is always held at the start of each loop iteration.

- We don't need to get a reference on the task structure, since the task
  is responsible for removing itself from the wait_list. There is no risk,
  like in the rwsem_down_read_failed() case, that a task would wake up and
  exit (thus destroying its task structure) while __rwsem_do_wake() is
  still running - wait_lock protects against that.
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Reviewed-by: Peter Hurley <peter@hurleysoftware.com>
Acked-by: Davidlohr Bueso <davidlohr.bueso@hp.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

When trying to acquire a read lock, the RWSEM_ACTIVE_READ_BIAS
adjustment doesn't cause other readers to block, so we never have to
worry about waking them back after canceling this adjustment in
rwsem_down_read_failed().

We also never want to steal the lock in rwsem_down_read_failed(), so we
don't have to grab the wait_lock either.
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Reviewed-by: Peter Hurley <peter@hurleysoftware.com>
Acked-by: Davidlohr Bueso <davidlohr.bueso@hp.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Remove the rwsem_down_failed_common function and replace it with two
identical copies of its code in rwsem_down_{read,write}_failed.

This is because we want to make different optimizations in
rwsem_down_{read,write}_failed; we are adding this pure-duplication
step as a separate commit in order to make it easier to check the
following steps.
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Reviewed-by: Peter Hurley <peter@hurleysoftware.com>
Acked-by: Davidlohr Bueso <davidlohr.bueso@hp.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

This change reduces the size of the spinlocked and TASK_UNINTERRUPTIBLE
sections in rwsem_down_failed_common():

- We only need the sem->wait_lock to insert ourselves on the wait_list;
  the waiter node can be prepared outside of the wait_lock.

- The task state only needs to be set to TASK_UNINTERRUPTIBLE immediately
  before checking if we actually need to sleep; it doesn't need to protect
  the entire function.
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Reviewed-by: Peter Hurley <peter@hurleysoftware.com>
Acked-by: Davidlohr Bueso <davidlohr.bueso@hp.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

We are not planning to add some new waiter flags, so we can convert the
waiter type into an enumeration.

Background: David Howells suggested I do this back when I tried adding
a new waiter type for unfair readers. However, I believe the cleanup
applies regardless of that use case.
Signed-off-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Reviewed-by: Peter Hurley <peter@hurleysoftware.com>
Acked-by: Davidlohr Bueso <davidlohr.bueso@hp.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Pull ARM SoC fix from Olof Johansson:
 "A late-arriving fix for musb on OMAP4, resolving an issue where the
  musb IP won't be clocked and thus not functional.  Small in scope,
  most of the lines changed is a longish comment."

* tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc:
  ARM: OMAP4: hwmod data: make 'ocp2scp_usb_phy_phy_48m" as the main clock

I think we could just move the full vm_iomap_memory() function into
util.h or similar, but I didn't get any reply from anybody actually
using nommu even to this trivial patch, so I'm not going to touch it any
more than required.

Here's the fairly minimal stub to make the nommu case at least
potentially work.  It doesn't seem like anybody cares, though.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Pull perf fix from Ingo Molnar:
 "This fix adds missing RCU read protection"

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  events: Protect access via task_subsys_state_check()

Merge tag 'omap-for-v3.9-rc6/fixes-signed' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap into fixes

From Tony Lindgren:
One MUSB regression fix that I forgot to send earlier. Without
this MUSB no longer works on omap4 based devices.

* tag 'omap-for-v3.9-rc6/fixes-signed' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap:
  ARM: OMAP4: hwmod data: make 'ocp2scp_usb_phy_phy_48m" as the main clock
Signed-off-by: Olof Johansson <olof@lixom.net>

Pull media fixes from Mauro Carvalho Chehab:
 "Two driver fixes.

  One avoids reading any file at a system with a cx25821 board
  (fortunately, this is not a common device).  The other one prevents
  reading after a buffer with ISDB-T devices based on mb86a20s."

* 'v4l_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media:
  [media] cx25821: do not expose broken video output streams
  [media] mb86a20s: Fix estimate_rate setting

Pull late parisc fixes from Helge Deller:
 "I know it's *very* late in the 3.9 release cycle, but since there
  aren't that many people testing the parisc linux kernel, a few (for
  our port) critical issues just showed up a few days back for the first
  time.

  What's in it?
   - add missing __ucmpdi2 symbol, which is required for btrfs on 32bit
     kernel.
   - change kunmap() macro to static inline function.  This fixes a
     debian/gcc-4.4 build error.
   - add locking when doing PTE updates.  This fixes random userspace
     crashes.
   - disable (optional) -mlong-calls compiler option for modules, else
     modules can't be loaded at runtime.
   - a smart patch by Will Deacon which fixes 64bit put_user() warnings
     on 32bit kernel."

* 'fixes-3.9-late' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux:
  parisc: use spin_lock_irqsave/spin_unlock_irqrestore for PTE updates
  parisc: disable -mlong-calls compiler option for kernel modules
  parisc: uaccess: fix compiler warnings caused by __put_user casting
  parisc: Change kunmap macro to static inline function
  parisc: Provide __ucmpdi2 to resolve undefined references in 32 bit builds.

variable_is_present() accesses '__efivars' directly, but when called via
gsmi_init() Michel reports observing the following crash,

  BUG: unable to handle kernel NULL pointer dereference at (null)
  IP: variable_is_present+0x55/0x170
  Call Trace:
    register_efivars+0x106/0x370
    gsmi_init+0x2ad/0x3da
    do_one_initcall+0x3f/0x170

The reason for the crash is that '__efivars' hasn't been initialised nor
has it been registered with register_efivars() by the time the google
EFI SMI driver runs.  The gsmi code uses its own struct efivars, and
therefore, a different variable list.  Fix the above crash by passing
the registered struct efivars to variable_is_present(), so that we
traverse the correct list.
Reported-by: Michel Lespinasse <walken@google.com>
Tested-by: Michel Lespinasse <walken@google.com>
Cc: Mike Waychison <mikew@google.com>
Cc: Matthew Garrett <matthew.garrett@nebula.com>
Cc: Seiji Aguchi <seiji.aguchi@hds.com>
Signed-off-by: Matt Fleming <matt.fleming@intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

In commit b0de59b5 ("TTY: do not update atime/mtime on read/write")
we removed timestamps from tty inodes to fix a security issue and waited
if something breaks.  Well, 'w', the utility to find out logged users
and their inactivity time broke.  It shows that users are inactive since
the time they logged in.

To revert to the old behaviour while still preventing attackers to
guess the password length, we update the timestamps in one-minute
intervals by this patch.
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

dprintk() shouldn't access @ring after it's unmapped.
Signed-off-by: Zhao Hongjiang <zhaohongjiang@huawei.com>
Cc: stable@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

 * The EFI variable anti-bricking algorithm merged in -rc8 broke booting
   on some Apple machines because they implement EFI spec 1.10, which
   doesn't provide a QueryVariableInfo() runtime function and the logic
   used to check for the existence of that function was insufficient.
   Fix from Josh Boyer.

 * The anti-bricking algorithm also introduced a compiler warning on
   32-bit. Fix from Borislav Petkov.
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>

User applications running on SMP kernels have long suffered from instability
and random segmentation faults.  This patch improves the situation although
there is more work to be done.

One of the problems is the various routines in pgtable.h that update page table
entries use different locking mechanisms, or no lock at all (set_pte_at).  This
change modifies the routines to all use the same lock pa_dbit_lock.  This lock
is used for dirty bit updates in the interruption code. The patch also purges
the TLB entries associated with the PTE to ensure that inconsistent values are
not used after the page table entry is updated.  The UP and SMP code are now
identical.

The change also includes a minor update to the purge_tlb_entries function in
cache.c to improve its efficiency.
Signed-off-by: John David Anglin <dave.anglin@bell.net>
Cc: Helge Deller <deller@gmx.de>
Signed-off-by: Helge Deller <deller@gmx.de>

CONFIG_MLONGCALLS was introduced in commit
ec758f98 to overcome linker issues when linking
huge linux kernels, e.g. with many modules linked in.

But in the kernel module loader there is no support yet for the new relocation
types, which is why modules built with -mlong-calls can't be loaded.
Furthermore, for modules long calls are not really necessary, since we already
use stub sections which resolve long distance calls.

So, let's just disable this compiler option when compiling kernel modules.
Signed-off-by: Helge Deller <deller@gmx.de>

When targetting 32-bit processors, __put_user emits a pair of stw
instructions for the 8-byte case. If the type of __val is a pointer, the
marshalling code casts it to the wider integer type of u64, resulting
in the following compiler warnings:

  kernel/signal.c: In function 'copy_siginfo_to_user':
  kernel/signal.c:2752:11: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
  kernel/signal.c:2752:11: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
  [...]

This patch fixes the warnings by removing the marshalling code and using
the correct output modifiers in the __put_{user,kernel}_asm64 macros
so that GCC will allocate the right registers without the need to
extract the two words explicitly.

Cc: Helge Deller <deller@gmx.de>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Helge Deller <deller@gmx.de>

Change kunmap macro to static inline function to fix build error
compiling drivers/base/dma-buf.c.

Without the change, the following error can occur:

   CC      drivers/base/dma-buf.o
drivers/base/dma-buf.c: In function 'dma_buf_kunmap':
drivers/base/dma-buf.c:427:46:
error: macro "kunmap" passed 3 arguments, but takes just 1

I believe parisc is the only arch to implement kunmap using a macro.
Signed-off-by: John David Anglin <dave.anglin@bell.net>
Cc: "James E.J. Bottomley" <jejb@parisc-linux.org>
Cc: Helge Deller <deller@gmx.de>
Signed-off-by: Helge Deller <deller@gmx.de>

The Debian experimental linux source package (3.8.5-1) build fails
with the following errors:
...
MODPOST 2016 modules
ERROR: "__ucmpdi2" [fs/btrfs/btrfs.ko] undefined!
ERROR: "__ucmpdi2" [drivers/md/dm-verity.ko] undefined!

The attached patch resolves this problem.  It is based on the s390
implementation of ucmpdi2.c.
Signed-off-by: John David Anglin <dave.anglin@bell.net>
Cc: "James E.J. Bottomley" <jejb@parisc-linux.org>
Signed-off-by: Helge Deller <deller@gmx.de>

Pull sparc fix from David Miller:
 "Brown paper bag fix for sparc64"

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc:
  sparc64: Fix missing put_cpu_var() in tlb_batch_add_one() when not batching.

Pull gpi fix from Linus Walleij:
 "This is a last minute revert for the GPIO tree, as Mike Dunn noticed
  breakage on some older PXA machines due to moving PXA GPIO initcalls
  to the module_init initlevel"

* tag 'gpio-v3.9-lastminute' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio:
  Revert "gpio: pxa: set initcall level to module init"

Reported-by: Meelis Roos <mroos@linux.ee>
Signed-off-by: David S. Miller <davem@davemloft.net>

This reverts commit 6c7e660a.

The commit causes breakage on several older PXA machines.
Reported-by: Mike Dunn <mikedunn@newsguy.com>
Acked-by: Haojian Zhuang <haojian.zhuang@linaro.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>

We need to check the runtime sys_table for the EFI version the firmware
specifies instead of just checking for a NULL QueryVariableInfo.  Older
implementations of EFI don't have QueryVariableInfo but the runtime is
a smaller structure, so the pointer to it may be pointing off into garbage.

This is apparently the case with several Apple firmwares that support EFI
1.10, and the current check causes them to no longer boot.  Fix based on
a suggestion from Matthew Garrett.
Signed-off-by: Josh Boyer <jwboyer@redhat.com>
Signed-off-by: Matt Fleming <matt.fleming@intel.com>

Fix this:

arch/x86/boot/compressed/eboot.c: In function ‘setup_efi_vars’:
arch/x86/boot/compressed/eboot.c:269:2: warning: passing argument 1 of ‘efi_call_phys’ makes pointer from integer without a cast [enabled by default]
In file included from arch/x86/boot/compressed/eboot.c:12:0:
/w/kernel/linux/arch/x86/include/asm/efi.h:8:33: note: expected ‘void *’ but argument is of type ‘long unsigned int’

after cc5a080c ("efi: Pass boot services variable info to runtime
code").
Reported-by: Paul Bolle <pebolle@tiscali.nl>
Cc: Matthew Garrett <matthew.garrett@nebula.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Matt Fleming <matt.fleming@intel.com>

Pull MIPS fix from Ralf Baechle:
 "Revert the change of the definition of PAGE_MASK which was prettier
  but broke a few relativly rare platforms"

* 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus:
  Revert "MIPS: page.h: Provide more readable definition for PAGE_MASK."

This reverts commit c17a6554.

Manuel Lauss writes:

lmo commit c17a6554 (MIPS: page.h: Provide more readable definition for
PAGE_MASK) apparently breaks ioremap of 36-bit addresses on my Alchemy
systems (PCI and PCMCIA) The reason is that in arch/mips/mm/ioremap.c
line 157  (phys_addr &= PAGE_MASK) bits 32-35 are cut off.  Seems the
new PAGE_MASK is explicitly 32bit, or one could make it signed instead
of unsigned long.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Pull crypto fix from Herbert Xu:
 "This fixes a kernel memory leak in the algif interface"

* git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
  crypto: algif - suppress sending source address information in recvmsg

Pull perf fixes from Ingo Molnar:
 "Misc fixes"

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf/x86: Fix offcore_rsp valid mask for SNB/IVB
  perf: Treat attr.config as u64 in perf_swevent_init()

I'm going to do an -rc8, so I'm just going to do this rather than delay
it any further. They are arguably stable material anyway.

* vm_ioremap_memory-examples:
  mtdchar: remove no-longer-used vma helpers
  vm: convert snd_pcm_lib_mmap_iomem() to vm_iomap_memory() helper
  vm: convert fb_mmap to vm_iomap_memory() helper
  vm: convert mtdchar mmap to vm_iomap_memory() helper
  vm: convert HPET mmap to vm_iomap_memory() helper