We generated a battery of 100 test cases from GSL taus113 implemention
and compare the results from a particular seed and a particular
iteration with our implementation in the kernel. We have verified on
32 and 64 bit machines that our taus113 kernel implementation gives
same results as GSL taus113 implementation:

  [    0.147370] prandom: seed boundary self test passed
  [    0.148078] prandom: 100 self tests passed

This is a Kconfig option that is disabled on default, just like the
crc32 init selftests in order to not unnecessary slow down boot process.
We also refactored out prandom_seed_very_weak() as it's now used in
multiple places in order to reduce redundant code.

GSL code we used for generating test cases:

  int i, j;
  srand(time(NULL));
  for (i = 0; i < 100; ++i) {
    int iteration = 500 + (rand() % 500);
    gsl_rng_default_seed = rand() + 1;
    gsl_rng *r = gsl_rng_alloc(gsl_rng_taus113);
    printf("\t{ %lu, ", gsl_rng_default_seed);
    for (j = 0; j < iteration - 1; ++j)
      gsl_rng_get(r);
    printf("%u, %lu },\n", iteration, gsl_rng_get(r));
    gsl_rng_free(r);
  }

Joint work with Hannes Frederic Sowa.

Cc: Florian Weimer <fweimer@redhat.com>
Cc: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: NDaniel Borkmann <dborkman@redhat.com>
Signed-off-by: NHannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Since we use prandom*() functions quite often in networking code
i.e. in UDP port selection, netfilter code, etc, upgrade the PRNG
from Pierre L'Ecuyer's original paper "Maximally Equidistributed
Combined Tausworthe Generators", Mathematics of Computation, 65,
213 (1996), 203--213 to the version published in his errata paper [1].

The Tausworthe generator is a maximally-equidistributed generator,
that is fast and has good statistical properties [1].

The version presented there upgrades the 3 state LFSR to a 4 state
LFSR with increased periodicity from about 2^88 to 2^113. The
algorithm is presented in [1] by the very same author who also
designed the original algorithm in [2].

Also, by increasing the state, we make it a bit harder for attackers
to "guess" the PRNGs internal state. See also discussion in [3].

Now, as we use this sort of weak initialization discussed in [3]
only between core_initcall() until late_initcall() time [*] for
prandom32*() users, namely in prandom_init(), it is less relevant
from late_initcall() onwards as we overwrite seeds through
prandom_reseed() anyways with a seed source of higher entropy, that
is, get_random_bytes(). In other words, a exhaustive keysearch of
96 bit would be needed. Now, with the help of this patch, this
state-search increases further to 128 bit. Initialization needs
to make sure that s1 > 1, s2 > 7, s3 > 15, s4 > 127.

taus88 and taus113 algorithm is also part of GSL. I added a test
case in the next patch to verify internal behaviour of this patch
with GSL and ran tests with the dieharder 3.31.1 RNG test suite:

$ dieharder -g 052 -a -m 10 -s 1 -S 4137730333 #taus88
$ dieharder -g 054 -a -m 10 -s 1 -S 4137730333 #taus113

With this seed configuration, in order to compare both, we get
the following differences:

algorithm                 taus88           taus113
rands/second [**]         1.61e+08         1.37e+08
sts_serial(4, 1st run)    WEAK             PASSED
sts_serial(9, 2nd run)    WEAK             PASSED
rgb_lagged_sum(31)        WEAK             PASSED

We took out diehard_sums test as according to the authors it is
considered broken and unusable [4]. Despite that and the slight
decrease in performance (which is acceptable), taus113 here passes
all 113 tests (only rgb_minimum_distance_5 in WEAK, the rest PASSED).
In general, taus/taus113 is considered "very good" by the authors
of dieharder [5].

The papers [1][2] states a single warm-up step is sufficient by
running quicktaus once on each state to ensure proper initialization
of ~s_{0}:

Our selection of (s) according to Table 1 of [1] row 1 holds the
condition L - k <= r - s, that is,

  (32 32 32 32) - (31 29 28 25) <= (25 27 15 22) - (18 2 7 13)

with r = k - q and q = (6 2 13 3) as also stated by the paper.
So according to [2] we are safe with one round of quicktaus for
initialization. However we decided to include the warm-up phase
of the PRNG as done in GSL in every case as a safety net. We also
use the warm up phase to make the output of the RNG easier to
verify by the GSL output.

In prandom_init(), we also mix random_get_entropy() into it, just
like drivers/char/random.c does it, jiffies ^ random_get_entropy().
random-get_entropy() is get_cycles(). xor is entropy preserving so
it is fine if it is not implemented by some architectures.

Note, this PRNG is *not* used for cryptography in the kernel, but
rather as a fast PRNG for various randomizations i.e. in the
networking code, or elsewhere for debugging purposes, for example.

[*]: In order to generate some "sort of pseduo-randomness", since
get_random_bytes() is not yet available for us, we use jiffies and
initialize states s1 - s3 with a simple linear congruential generator
(LCG), that is x <- x * 69069; and derive s2, s3, from the 32bit
initialization from s1. So the above quote from [3] accounts only
for the time from core to late initcall, not afterwards.
[**] Single threaded run on MacBook Air w/ Intel Core i5-3317U

 [1] http://www.iro.umontreal.ca/~lecuyer/myftp/papers/tausme2.ps
 [2] http://www.iro.umontreal.ca/~lecuyer/myftp/papers/tausme.ps
 [3] http://thread.gmane.org/gmane.comp.encryption.general/12103/
 [4] http://code.google.com/p/dieharder/source/browse/trunk/libdieharder/diehard_sums.c?spec=svn490&r=490#20
 [5] http://www.phy.duke.edu/~rgb/General/dieharder.php

Joint work with Hannes Frederic Sowa.

Cc: Florian Weimer <fweimer@redhat.com>
Cc: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: NDaniel Borkmann <dborkman@redhat.com>
Signed-off-by: NHannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The Tausworthe PRNG is initialized at late_initcall time. At that time the
entropy pool serving get_random_bytes is not filled sufficiently. This
patch adds an additional reseeding step as soon as the nonblocking pool
gets marked as initialized.

On some machines it might be possible that late_initcall gets called after
the pool has been initialized. In this situation we won't reseed again.

(A call to prandom_seed_late blocks later invocations of early reseed
attempts.)

Joint work with Daniel Borkmann.

Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: NHannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: NDaniel Borkmann <dborkman@redhat.com>
Acked-by: N"Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The current Tausworthe PRNG is never reseeded with truly random data after
the first attempt in late_initcall. As this PRNG is used for some critical
random data as e.g. UDP port randomization we should try better and reseed
the PRNG once in a while with truly random data from get_random_bytes().

When we reseed with prandom_seed we now make also sure to throw the first
output away. This suffices the reseeding procedure.

The delay calculation is based on a proposal from Eric Dumazet.

Joint work with Daniel Borkmann.

Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: NHannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: NDaniel Borkmann <dborkman@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

For properly initialising the Tausworthe generator [1], we have
a strict seeding requirement, that is, s1 > 1, s2 > 7, s3 > 15.

Commit 697f8d03 ("random32: seeding improvement") introduced
a __seed() function that imposes boundary checks proposed by the
errata paper [2] to properly ensure above conditions.

However, we're off by one, as the function is implemented as:
"return (x < m) ? x + m : x;", and called with __seed(X, 1),
__seed(X, 7), __seed(X, 15). Thus, an unwanted seed of 1, 7, 15
would be possible, whereas the lower boundary should actually
be of at least 2, 8, 16, just as GSL does. Fix this, as otherwise
an initialization with an unwanted seed could have the effect
that Tausworthe's PRNG properties cannot not be ensured.

Note that this PRNG is *not* used for cryptography in the kernel.

 [1] http://www.iro.umontreal.ca/~lecuyer/myftp/papers/tausme.ps
 [2] http://www.iro.umontreal.ca/~lecuyer/myftp/papers/tausme2.ps

Joint work with Hannes Frederic Sowa.

Fixes: 697f8d03 ("random32: seeding improvement")
Cc: Stephen Hemminger <stephen@networkplumber.org>
Cc: Florian Weimer <fweimer@redhat.com>
Cc: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: NDaniel Borkmann <dborkman@redhat.com>
Signed-off-by: NHannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We can safely reduce the number of test cases by a tenth.
There is no particular need to run as many as we're running
now for crc32{,c}_combine, that gives us still ~8000 tests
we're doing if people run kernels with crc selftests enabled
which is perfectly fine.
Signed-off-by: NDaniel Borkmann <dborkman@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Fengguang reports that when crc32 selftests are running on startup, on
some e.g. 32bit systems, we can get a CPU stall like "INFO: rcu_sched
self-detected stall on CPU { 0} (t=2101 jiffies g=4294967081 c=4294967080
q=41)". As this is not intended, add a cond_resched() at the end of a
test case to fix it. Introduced by efba721f ("lib: crc32: add test cases
for crc32{, c}_combine routines").
Reported-by: NFengguang Wu <fengguang.wu@intel.com>
Signed-off-by: NDaniel Borkmann <dborkman@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We already have 100 test cases for crcs itself, so split the test
buffer with a-prio known checksums, and test crc of two blocks
against crc of the whole block for the same results.

Output/result with CONFIG_CRC32_SELFTEST=y:

  [    2.687095] crc32: CRC_LE_BITS = 64, CRC_BE BITS = 64
  [    2.687097] crc32: self tests passed, processed 225944 bytes in 278177 nsec
  [    2.687383] crc32c: CRC_LE_BITS = 64
  [    2.687385] crc32c: self tests passed, processed 225944 bytes in 141708 nsec
  [    7.336771] crc32_combine: 113072 self tests passed
  [   12.050479] crc32c_combine: 113072 self tests passed
  [   17.633089] alg: No test for crc32 (crc32-pclmul)
Signed-off-by: NDaniel Borkmann <dborkman@redhat.com>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch adds a combinator to merge two or more crc32{,c}s
into a new one. This is useful for checksum computations of
fragmented skbs that use crc32/crc32c as checksums.

The arithmetics for combining both in the GF(2) was taken and
slightly modified from zlib. Only passing two crcs is insufficient
as two crcs and the length of the second piece is needed for
merging. The code is made generic, so that only polynomials
need to be passed for crc32_le resp. crc32c_le.
Signed-off-by: NDaniel Borkmann <dborkman@redhat.com>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This is nothing more but a whitepace cleanup, as 80 chars is not a
hard but soft limit, and otherwise makes the test cases array really
look ugly. So fix it up.
Signed-off-by: NDaniel Borkmann <dborkman@redhat.com>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Commit b1adaf65 ("[SCSI] block: add sg buffer copy helper
functions") introduces two sg buffer copy helpers, and calls
flush_kernel_dcache_page() on pages in SG list after these pages are
written to.

Unfortunately, the commit may introduce a potential bug:

 - Before sending some SCSI commands, kmalloc() buffer may be passed to
   block layper, so flush_kernel_dcache_page() can see a slab page
   finally

 - According to cachetlb.txt, flush_kernel_dcache_page() is only called
   on "a user page", which surely can't be a slab page.

 - ARCH's implementation of flush_kernel_dcache_page() may use page
   mapping information to do optimization so page_mapping() will see the
   slab page, then VM_BUG_ON() is triggered.

Aaro Koskinen reported the bug on ARM/kirkwood when DEBUG_VM is enabled,
and this patch fixes the bug by adding test of '!PageSlab(miter->page)'
before calling flush_kernel_dcache_page().
Signed-off-by: NMing Lei <ming.lei@canonical.com>
Reported-by: NAaro Koskinen <aaro.koskinen@iki.fi>
Tested-by: NSimon Baatz <gmbnomis@gmail.com>
Cc: Russell King - ARM Linux <linux@arm.linux.org.uk>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Aaro Koskinen <aaro.koskinen@iki.fi>
Acked-by: NCatalin Marinas <catalin.marinas@arm.com>
Cc: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Cc: Tejun Heo <tj@kernel.org>
Cc: "James E.J. Bottomley" <JBottomley@parallels.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: <stable@vger.kernel.org>	[3.2+]
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Without the timer debugging, the delayed kobject release will just
result in undebuggable oopses if it triggers any latent bugs.  That
doesn't actually help debugging at all.

So make DEBUG_KOBJECT_RELEASE depend on DEBUG_OBJECTS_TIMERS to avoid
having people enable one without the other.
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Export the interface to be used within modules.
Signed-off-by: NMatias Bjorling <m@bjorling.me>
Acked-by: NTejun Heo <tj@kernel.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Useful for locating buggy drivers on kernel oops.

It may add dozens of new lines to boot dmesg. DEBUG_KOBJECT_RELEASE is
hopefully only enabled in debug kernels (like maybe the Fedora rawhide
one, or at developers), so being a bit more verbose is likely ok.
Acked-by: NRussell King <rmk+kernel@arm.linux.org.uk>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: NFengguang Wu <fengguang.wu@intel.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Make use of arch_mutex_cpu_relax() so architectures can override the
default cpu_relax() semantics.
This is especially useful for s390, where cpu_relax() means that we
yield() the current (virtual) cpu and therefore is very expensive,
and would contradict the whole purpose of the lockless cmpxchg loop.
Signed-off-by: NHeiko Carstens <heiko.carstens@de.ibm.com>

In kobj_ns_current_may_mount the default should be to allow the mount.
The test is only for a single kobj_ns_type at a time, and unless there
is a reason to prevent it the mounting sysfs should be allowed.
Subsystems that are not registered can't have are not involved so can't
have a reason to prevent mounting sysfs.

This is a bug-fix to commit 7dc5dbc8 ("sysfs: Restrict mounting
sysfs") that came in via the userns tree during the 3.12 merge window.
Reported-and-tested-by: NJames Hogan <james.hogan@imgtec.com>
Signed-off-by: N"Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The 64-bit cmpxchg operation on the lockref is ordered by virtue of
hazarding between the cmpxchg operation and the reference count
manipulation. On weakly ordered memory architectures (such as ARM), it
can be of great benefit to omit the barrier instructions where they are
not needed.

This patch moves the lockless lockref code over to a cmpxchg64_relaxed
operation, which doesn't provide barrier semantics. If the operation
isn't defined, we simply #define it as the usual 64-bit cmpxchg macro.

Cc: Waiman Long <Waiman.Long@hp.com>
Signed-off-by: NWill Deacon <will.deacon@arm.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

In kobj_ns_current_may_mount the default should be to allow the
mount.  The test is only for a single kobj_ns_type at a time, and unless
there is a reason to prevent it the mounting sysfs should be allowed.
Subsystems that are not registered can't have are not involved so can't
have a reason to prevent mounting sysfs.

This is a bug-fix to:
    commit 7dc5dbc8
    Author: Eric W. Biederman <ebiederm@xmission.com>
    Date:   Mon Mar 25 20:07:01 2013 -0700

        sysfs: Restrict mounting sysfs

        Don't allow mounting sysfs unless the caller has CAP_SYS_ADMIN rights
        over the net namespace.  The principle here is if you create or have
        capabilities over it you can mount it, otherwise you get to live with
        what other people have mounted.

        Instead of testing this with a straight forward ns_capable call,
        perform this check the long and torturous way with kobject helpers,
        this keeps direct knowledge of namespaces out of sysfs, and preserves
        the existing sysfs abstractions.
Acked-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: N"Eric W. Biederman" <ebiederm@xmission.com>

That came in via the userns tree during the 3.12 merge window.
Reported-by: NJames Hogan <james.hogan@imgtec.com>
Signed-off-by: N"Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

To be able to use the hex ascii functions in case sensitive environments
the array hex_asc_upper[] and the needed functions for hex_byte_pack_upper()
are introduced.
Signed-off-by: NAndre Naujoks <nautsch2@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The cmpxchg() function tends not to support 64-bit arguments on 32-bit
architectures.  This could be either due to use of unsigned long
arguments (like on ARM) or lack of instruction support (cmpxchgq on
x86).  However, these architectures may implement a specific cmpxchg64()
function to provide 64-bit cmpxchg support instead.

Since the lockref code requires a 64-bit cmpxchg and relies on the
architecture selecting ARCH_USE_CMPXCHG_LOCKREF, move to using cmpxchg64
instead of cmpxchg and allow 32-bit architectures to make use of the
lockless lockref implementation.

Cc: Waiman Long <Waiman.Long@hp.com>
Signed-off-by: NWill Deacon <will.deacon@arm.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

After the last architecture switched to generic hard irqs the config
options HAVE_GENERIC_HARDIRQS & GENERIC_HARDIRQS and the related code
for !CONFIG_GENERIC_HARDIRQS can be removed.
Signed-off-by: NMartin Schwidefsky <schwidefsky@de.ibm.com>

Unfortunately, even with a softdep some distros fail to include
the necessary modules in the initrd.  Therefore this patch adds
a fallback path to restore existing behaviour where we cannot
load the new crypto crct10dif algorithm.

In order to do this, the underlying crct10dif has been split out
from the crypto implementation so that it can be used on the
fallback path.
Signed-off-by: NHerbert Xu <herbert@gondor.apana.org.au>

LZ4 compression and decompression functions require different in
signedness input/output parameters: unsigned char for compression and
signed char for decompression.

Change decompression API to require "(const) unsigned char *".
Signed-off-by: NSergey Senozhatsky <sergey.senozhatsky@gmail.com>
Cc: Kyungsik Lee <kyungsik.lee@lge.com>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Yann Collet <yann.collet.73@gmail.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

With users of radix_tree_preload() run from interrupt (block/blk-ioc.c is
one such possible user), the following race can happen:

radix_tree_preload()
...
radix_tree_insert()
  radix_tree_node_alloc()
    if (rtp->nr) {
      ret = rtp->nodes[rtp->nr - 1];
<interrupt>
...
radix_tree_preload()
...
radix_tree_insert()
  radix_tree_node_alloc()
    if (rtp->nr) {
      ret = rtp->nodes[rtp->nr - 1];

And we give out one radix tree node twice.  That clearly results in radix
tree corruption with different results (usually OOPS) depending on which
two users of radix tree race.

We fix the problem by making radix_tree_node_alloc() always allocate fresh
radix tree nodes when in interrupt.  Using preloading when in interrupt
doesn't make sense since all the allocations have to be atomic anyway and
we cannot steal nodes from process-context users because some users rely
on radix_tree_insert() succeeding after radix_tree_preload().
in_interrupt() check is somewhat ugly but we cannot simply key off passed
gfp_mask as that is acquired from root_gfp_mask() and thus the same for
all preload users.

Another part of the fix is to avoid node preallocation in
radix_tree_preload() when passed gfp_mask doesn't allow waiting.  Again,
preallocation in such case doesn't make sense and when preallocation would
happen in interrupt we could possibly leak some allocated nodes.  However,
some users of radix_tree_preload() require following radix_tree_insert()
to succeed.  To avoid unexpected effects for these users,
radix_tree_preload() only warns if passed gfp mask doesn't allow waiting
and we provide a new function radix_tree_maybe_preload() for those users
which get different gfp mask from different call sites and which are
prepared to handle radix_tree_insert() failure.
Signed-off-by: NJan Kara <jack@suse.cz>
Cc: Jens Axboe <jaxboe@fusionio.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

No reason require rbtree test code to be a module, allow it to be builtin
(streamlines my development process)
Signed-off-by: NCody P Schafer <cody@linux.vnet.ibm.com>
Reviewed-by: NSeth Jennings <sjenning@linux.vnet.ibm.com>
Cc: David Woodhouse <David.Woodhouse@intel.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Michel Lespinasse <walken@google.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Just check that we examine all nodes in the tree for the postorder
iteration.
Signed-off-by: NCody P Schafer <cody@linux.vnet.ibm.com>
Reviewed-by: NSeth Jennings <sjenning@linux.vnet.ibm.com>
Cc: David Woodhouse <David.Woodhouse@intel.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Michel Lespinasse <walken@google.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Postorder iteration yields all of a node's children prior to yielding the
node itself, and this particular implementation also avoids examining the
leaf links in a node after that node has been yielded.

In what I expect will be its most common usage, postorder iteration allows
the deletion of every node in an rbtree without modifying the rbtree nodes
(no _requirement_ that they be nulled) while avoiding referencing child
nodes after they have been "deleted" (most commonly, freed).

I have only updated zswap to use this functionality at this point, but
numerous bits of code (most notably in the filesystem drivers) use a hand
rolled postorder iteration that NULLs child links as it traverses the
tree.  Each of those instances could be replaced with this common
implementation.

1 & 2 add rbtree postorder iteration functions.
3 adds testing of the iteration to the rbtree runtime tests
4 allows building the rbtree runtime tests as builtins
5 updates zswap.

This patch:

Add postorder iteration functions for rbtree.  These are useful for safely
freeing an entire rbtree without modifying the tree at all.
Signed-off-by: NCody P Schafer <cody@linux.vnet.ibm.com>
Reviewed-by: NSeth Jennings <sjenning@linux.vnet.ibm.com>
Cc: David Woodhouse <David.Woodhouse@intel.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Michel Lespinasse <walken@google.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

When decompressing into memory, the output buffer length is set to some
arbitrarily high value (0x7fffffff) to indicate the output is, virtually,
unlimited in size.

The problem with this is that some platforms have their physical memory at
high physical addresses (0x80000000 or more), and that the output buffer
address and its "unlimited" length cannot be added without overflowing.
An example of this can be found in inflate_fast():

/* next_out is the output buffer address */
out = strm->next_out - OFF;
/* avail_out is the output buffer size. end will overflow if the output
 * address is >= 0x80000104 */
end = out + (strm->avail_out - 257);

This has huge consequences on the performance of kernel decompression,
since the following exit condition of inflate_fast() will be always true:

} while (in < last && out < end);

Indeed, "end" has overflowed and is now always lower than "out".  As a
result, inflate_fast() will return after processing one single byte of
input data, and will thus need to be called an unreasonably high number of
times.  This probably went unnoticed because kernel decompression is fast
enough even with this issue.

Nonetheless, adjusting the output buffer length in such a way that the
above pointer arithmetic never overflows results in a kernel decompression
that is about 3 times faster on affected machines.
Signed-off-by: NAlexandre Courbot <acourbot@nvidia.com>
Tested-by: NJon Medhurst <tixy@linaro.org>
Cc: Stephen Warren <swarren@wwwdotorg.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: NGu Zheng <guz.fnst@cn.fujitsu.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The documentation mentions a "name" parameter, which does not exist.  This
commit removes such mention from the function documentation.
Signed-off-by: NEmilio López <emilio@elopez.com.ar>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Use the helper function instead of __GFP_ZERO.
Signed-off-by: NJoe Perches <joe@perches.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

In struct gen_pool_chunk, end_addr means the end address of memory chunk
(inclusive), but in the implementation it is treated as address + size of
memory chunk (exclusive), so it points to the address plus one instead of
correct ending address.

The ending address of memory chunk plus one will cause overflow on the
memory chunk including the last address of memory map, e.g.  when starting
address is 0xFFF00000 and size is 0x100000 on 32bit machine, ending
address will be 0x100000000.

Use correct ending address like starting address + size - 1.

[akpm@linux-foundation.org: add comment to struct gen_pool_chunk:end_addr]
Signed-off-by: NJoonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Percpu frontend for allocating ids. With percpu allocation (that works),
it's impossible to guarantee it will always be possible to allocate all
nr_tags - typically, some will be stuck on a remote percpu freelist
where the current job can't get to them.

We do guarantee that it will always be possible to allocate at least
(nr_tags / 2) tags - this is done by keeping track of which and how many
cpus have tags on their percpu freelists. On allocation failure if
enough cpus have tags that there could potentially be (nr_tags / 2) tags
stuck on remote percpu freelists, we then pick a remote cpu at random to
steal from.

Note that there's no cpu hotplug notifier - we don't care, because
steal_tags() will eventually get the down cpu's tags. We _could_ satisfy
more allocations if we had a notifier - but we'll still meet our
guarantees and it's absolutely not a correctness issue, so I don't think
it's worth the extra code.

From akpm:

    "It looks OK to me (that's as close as I get to an ack :))

v6 changes:
  - Add #include <linux/cpumask.h> to include/linux/percpu_ida.h to
    make alpha/arc builds happy (Fengguang)
  - Move second (cpu >= nr_cpu_ids) check inside of first check scope
    in steal_tags() (akpm + nab)

v5 changes:
  - Change percpu_ida->cpus_have_tags to cpumask_t (kmo + akpm)
  - Add comment for percpu_ida_cpu->lock + ->nr_free (kmo + akpm)
  - Convert steal_tags() to use cpumask_weight() + cpumask_next() +
    cpumask_first() + cpumask_clear_cpu() (kmo + akpm)
  - Add comment for alloc_global_tags() (kmo + akpm)
  - Convert percpu_ida_alloc() to use cpumask_set_cpu() (kmo + akpm)
  - Convert percpu_ida_free() to use cpumask_set_cpu() (kmo + akpm)
  - Drop percpu_ida->cpus_have_tags allocation in percpu_ida_init()
    (kmo + akpm)
  - Drop percpu_ida->cpus_have_tags kfree in percpu_ida_destroy()
    (kmo + akpm)
  - Add comment for percpu_ida_alloc @ gfp (kmo + akpm)
  - Move to percpu_ida.c + percpu_ida.h (kmo + akpm + nab)

v4 changes:

  - Fix tags.c reference in percpu_ida_init (akpm)
Signed-off-by: NKent Overstreet <kmo@daterainc.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: "Nicholas A. Bellinger" <nab@linux-iscsi.org>
Signed-off-by: NNicholas Bellinger <nab@linux-iscsi.org>

The only actual current lockref user (dcache) uses zero reference counts
even for perfectly live dentries, because it's a cache: there may not be
any users, but that doesn't mean that we want to throw away the dentry.

At the same time, the dentry cache does have a notion of a truly "dead"
dentry that we must not even increment the reference count of, because
we have pruned it and it is not valid.

Currently that distinction is not visible in the lockref itself, and the
dentry cache validation uses "lockref_get_or_lock()" to either get a new
reference to a dentry that already had existing references (and thus
cannot be dead), or get the dentry lock so that we can then verify the
dentry and increment the reference count under the lock if that
verification was successful.

That's all somewhat complicated.

This adds the concept of being "dead" to the lockref itself, by simply
using a count that is negative.  This allows a usage scenario where we
can increment the refcount of a dentry without having to validate it,
and pushing the special "we killed it" case into the lockref code.

The dentry code itself doesn't actually use this yet, and it's probably
too late in the merge window to do that code (the dentry_kill() code
with its "should I decrement the count" logic really is pretty complex
code), but let's introduce the concept at the lockref level now.
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The code got rewritten, but the comments got copied as-is from older
versions, and as a result the argument name in the comment didn't
actually match the code any more.
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

This patch reinstates commits
	67822649
	39761214
	0b95a7f8
	31d93962
	2d31e518

Now that module softdeps are in the kernel we can use that to resolve
the boot issue which cause the revert.
Signed-off-by: NHerbert Xu <herbert@gondor.apana.org.au>

Frame pointer on ARC doesn't serve the conventional purpose of stack
unwinding due to the typical way ABI designates it's usage.
Thus it's explicit usage on ARC is discouraged (gcc is free to use it,
for some tricky stack frames even if -fomit-frame-pointer).

Hence no point enabling it for ARC.

References: http://www.spinics.net/lists/kernel/msg1593937.htmlSigned-off-by: NVineet Gupta <vgupta@synopsys.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: "Paul E. McKenney" <paul.mckenney@linaro.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Michel Lespinasse <walken@google.com>
Cc: linux-kernel@vger.kernel.org

New formats: %p[dD][234]?.  The next pointer is interpreted as struct dentry *
or struct file * resp. ('d' => dentry, 'D' => file) and the last component(s)
of pathname are printed (%pd => just the last one, %pd2 => the last two, etc.)
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

While we are likley to succeed and break out of this loop, it isn't
guaranteed.  We should be power and thread friendly if we do have to
go around for a second (or third, or more) attempt.
Signed-off-by: NTony Luck <tony.luck@intel.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Instead of taking the spinlock, the lockless versions atomically check
that the lock is not taken, and do the reference count update using a
cmpxchg() loop.  This is semantically identical to doing the reference
count update protected by the lock, but avoids the "wait for lock"
contention that you get when accesses to the reference count are
contended.

Note that a "lockref" is absolutely _not_ equivalent to an atomic_t.
Even when the lockref reference counts are updated atomically with
cmpxchg, the fact that they also verify the state of the spinlock means
that the lockless updates can never happen while somebody else holds the
spinlock.

So while "lockref_put_or_lock()" looks a lot like just another name for
"atomic_dec_and_lock()", and both optimize to lockless updates, they are
fundamentally different: the decrement done by atomic_dec_and_lock() is
truly independent of any lock (as long as it doesn't decrement to zero),
so a locked region can still see the count change.

The lockref structure, in contrast, really is a *locked* reference
count.  If you hold the spinlock, the reference count will be stable and
you can modify the reference count without using atomics, because even
the lockless updates will see and respect the state of the lock.

In order to enable the cmpxchg lockless code, the architecture needs to
do three things:

 (1) Make sure that the "arch_spinlock_t" and an "unsigned int" can fit
     in an aligned u64, and have a "cmpxchg()" implementation that works
     on such a u64 data type.

 (2) define a helper function to test for a spinlock being unlocked
     ("arch_spin_value_unlocked()")

 (3) select the "ARCH_USE_CMPXCHG_LOCKREF" config variable in its
     Kconfig file.

This enables it for x86-64 (but not 32-bit, we'd need to make sure
cmpxchg() turns into the proper cmpxchg8b in order to enable it for
32-bit mode).
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>