Skip to content

cloud-kernel
cloud-kernel

openanolis / cloud-kernel
1 年多 前同步成功

通知 160
Star 36
Fork 7
cloud-kernel
cloud-kernel
切换分支/标签
  1. 11 1月, 2017 1 次提交
  3. 06 1月, 2017 3 次提交
  5. 30 12月, 2016 1 次提交
    • L
      mm: optimize PageWaiters bit use for unlock_page() · b91e1302
      由 Linus Torvalds 提交于 
      In commit 62906027 ("mm: add PageWaiters indicating tasks are
waiting for a page bit") Nick Piggin made our page locking no longer
unconditionally touch the hashed page waitqueue, which not only helps
performance in general, but is particularly helpful on NUMA machines
where the hashed wait queues can bounce around a lot.

However, the "clear lock bit atomically and then test the waiters bit"
sequence turns out to be much more expensive than it needs to be,
because you get a nasty stall when trying to access the same word that
just got updated atomically.

On architectures where locking is done with LL/SC, this would be trivial
to fix with a new primitive that clears one bit and tests another
atomically, but that ends up not working on x86, where the only atomic
operations that return the result end up being cmpxchg and xadd.  The
atomic bit operations return the old value of the same bit we changed,
not the value of an unrelated bit.

On x86, we could put the lock bit in the high bit of the byte, and use
"xadd" with that bit (where the overflow ends up not touching other
bits), and look at the other bits of the result.  However, an even
simpler model is to just use a regular atomic "and" to clear the lock
bit, and then the sign bit in eflags will indicate the resulting state
of the unrelated bit #7.

So by moving the PageWaiters bit up to bit #7, we can atomically clear
the lock bit and test the waiters bit on x86 too.  And architectures
with LL/SC (which is all the usual RISC suspects), the particular bit
doesn't matter, so they are fine with this approach too.

This avoids the extra access to the same atomic word, and thus avoids
the costly stall at page unlock time.

The only downside is that the interface ends up being a bit odd and
specialized: clear a bit in a byte, and test the sign bit.  Nick doesn't
love the resulting name of the new primitive, but I'd rather make the
name be descriptive and very clear about the limitation imposed by
trying to work across all relevant architectures than make it be some
generic thing that doesn't make the odd semantics explicit.

So this introduces the new architecture primitive

    clear_bit_unlock_is_negative_byte();

and adds the trivial implementation for x86.  We have a generic
non-optimized fallback (that just does a "clear_bit()"+"test_bit(7)"
combination) which can be overridden by any architecture that can do
better.  According to Nick, Power has the same hickup x86 has, for
example, but some other architectures may not even care.

All these optimizations mean that my page locking stress-test (which is
just executing a lot of small short-lived shell scripts: "make test" in
the git source tree) no longer makes our page locking look horribly bad.
Before all these optimizations, just the unlock_page() costs were just
over 3% of all CPU overhead on "make test".  After this, it's down to
0.66%, so just a quarter of the cost it used to be.

(The difference on NUMA is bigger, but there this micro-optimization is
likely less noticeable, since the big issue on NUMA was not the accesses
to 'struct page', but the waitqueue accesses that were already removed
by Nick's earlier commit).
Acked-by: NNick Piggin <npiggin@gmail.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Bob Peterson <rpeterso@redhat.com>
Cc: Steven Whitehouse <swhiteho@redhat.com>
Cc: Andrew Lutomirski <luto@kernel.org>
Cc: Andreas Gruenbacher <agruenba@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
      b91e1302
  7. 28 12月, 2016 2 次提交
  9. 27 12月, 2016 1 次提交
    • J
      mm: Invalidate DAX radix tree entries only if appropriate · c6dcf52c
      由 Jan Kara 提交于 
      Currently invalidate_inode_pages2_range() and invalidate_mapping_pages()
just delete all exceptional radix tree entries they find. For DAX this
is not desirable as we track cache dirtiness in these entries and when
they are evicted, we may not flush caches although it is necessary. This
can for example manifest when we write to the same block both via mmap
and via write(2) (to different offsets) and fsync(2) then does not
properly flush CPU caches when modification via write(2) was the last
one.

Create appropriate DAX functions to handle invalidation of DAX entries
for invalidate_inode_pages2_range() and invalidate_mapping_pages() and
wire them up into the corresponding mm functions.
Acked-by: NJohannes Weiner <hannes@cmpxchg.org>
Reviewed-by: NRoss Zwisler <ross.zwisler@linux.intel.com>
Signed-off-by: NJan Kara <jack@suse.cz>
Signed-off-by: NDan Williams <dan.j.williams@intel.com>
      c6dcf52c
  11. 26 12月, 2016 5 次提交
    • N
      mm: add PageWaiters indicating tasks are waiting for a page bit · 62906027
      由 Nicholas Piggin 提交于 
      Add a new page flag, PageWaiters, to indicate the page waitqueue has
tasks waiting. This can be tested rather than testing waitqueue_active
which requires another cacheline load.

This bit is always set when the page has tasks on page_waitqueue(page),
and is set and cleared under the waitqueue lock. It may be set when
there are no tasks on the waitqueue, which will cause a harmless extra
wakeup check that will clears the bit.

The generic bit-waitqueue infrastructure is no longer used for pages.
Instead, waitqueues are used directly with a custom key type. The
generic code was not flexible enough to have PageWaiters manipulation
under the waitqueue lock (which simplifies concurrency).

This improves the performance of page lock intensive microbenchmarks by
2-3%.

Putting two bits in the same word opens the opportunity to remove the
memory barrier between clearing the lock bit and testing the waiters
bit, after some work on the arch primitives (e.g., ensuring memory
operand widths match and cover both bits).
Signed-off-by: NNicholas Piggin <npiggin@gmail.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Bob Peterson <rpeterso@redhat.com>
Cc: Steven Whitehouse <swhiteho@redhat.com>
Cc: Andrew Lutomirski <luto@kernel.org>
Cc: Andreas Gruenbacher <agruenba@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
      62906027
    • N
      mm: Use owner_priv bit for PageSwapCache, valid when PageSwapBacked · 6326fec1
      由 Nicholas Piggin 提交于 
      A page is not added to the swap cache without being swap backed,
so PageSwapBacked mappings can use PG_owner_priv_1 for PageSwapCache.
Signed-off-by: NNicholas Piggin <npiggin@gmail.com>
Acked-by: NHugh Dickins <hughd@google.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Bob Peterson <rpeterso@redhat.com>
Cc: Steven Whitehouse <swhiteho@redhat.com>
Cc: Andrew Lutomirski <luto@kernel.org>
Cc: Andreas Gruenbacher <agruenba@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
      6326fec1
    • T
      ktime: Get rid of ktime_equal() · 1f3a8e49
      由 Thomas Gleixner 提交于 
      No point in going through loops and hoops instead of just comparing the
values.
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
      1f3a8e49
    • T
      ktime: Cleanup ktime_set() usage · 8b0e1953
      由 Thomas Gleixner 提交于 
      ktime_set(S,N) was required for the timespec storage type and is still
useful for situations where a Seconds and Nanoseconds part of a time value
needs to be converted. For anything where the Seconds argument is 0, this
is pointless and can be replaced with a simple assignment.
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
      8b0e1953
    • T
      ktime: Get rid of the union · 2456e855
      由 Thomas Gleixner 提交于 
      ktime is a union because the initial implementation stored the time in
scalar nanoseconds on 64 bit machine and in a endianess optimized timespec
variant for 32bit machines. The Y2038 cleanup removed the timespec variant
and switched everything to scalar nanoseconds. The union remained, but
become completely pointless.

Get rid of the union and just keep ktime_t as simple typedef of type s64.

The conversion was done with coccinelle and some manual mopping up.
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
      2456e855
  13. 25 12月, 2016 9 次提交
  15. 24 12月, 2016 1 次提交
  17. 23 12月, 2016 2 次提交
  19. 21 12月, 2016 5 次提交
    • L
      ACPI / osl: Remove deprecated acpi_get_table_with_size()/early_acpi_os_unmap_memory() · 8d3523fb
      由 Lv Zheng 提交于 
      Since all users are cleaned up, remove the 2 deprecated APIs due to no
users.
As a Linux variable rather than an ACPICA variable, acpi_gbl_permanent_mmap
is renamed to acpi_permanent_mmap to have a consistent coding style across
entire Linux ACPI subsystem.
Signed-off-by: NLv Zheng <lv.zheng@intel.com>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>
      8d3523fb
    • L
      ACPICA: Tables: Back port acpi_get_table_with_size() and... · 174cc718
      由 Lv Zheng 提交于 
      ACPICA: Tables: Back port acpi_get_table_with_size() and early_acpi_os_unmap_memory() from Linux kernel

ACPICA commit cac6790954d4d752a083e6122220b8a22febcd07

This patch back ports Linux acpi_get_table_with_size() and
early_acpi_os_unmap_memory() into ACPICA upstream to reduce divergences.

The 2 APIs are used by Linux as table management APIs for long time, it
contains a hidden logic that during the early stage, the mapped tables
should be unmapped before the early stage ends.

During the early stage, tables are handled by the following sequence:
 acpi_get_table_with_size();
 parse the table
 early_acpi_os_unmap_memory();
During the late stage, tables are handled by the following sequence:
 acpi_get_table();
 parse the table
Linux uses acpi_gbl_permanent_mmap to distinguish the early stage and the
late stage.

The reasoning of introducing acpi_get_table_with_size() is: ACPICA will
remember the early mapped pointer in acpi_get_table() and Linux isn't able to
prevent ACPICA from using the wrong early mapped pointer during the late
stage as there is no API provided from ACPICA to be an inverse of
acpi_get_table() to forget the early mapped pointer.

But how ACPICA can work with the early/late stage requirement? Inside of
ACPICA, tables are ensured to be remained in "INSTALLED" state during the
early stage, and they are carefully not transitioned to "VALIDATED" state
until the late stage. So the same logic is in fact implemented inside of
ACPICA in a different way. The gap is only that the feature is not provided
to the OSPMs in an accessible external API style.

It then is possible to fix the gap by providing an inverse of
acpi_get_table() from ACPICA, so that the two Linux sequences can be
combined:
 acpi_get_table();
 parse the table
 acpi_put_table();
In order to work easier with the current Linux code, acpi_get_table() and
acpi_put_table() is implemented in a usage counting based style:
 1. When the usage count of the table is increased from 0 to 1, table is
    mapped and .Pointer is set with the mapping address (VALIDATED);
 2. When the usage count of the table is decreased from 1 to 0, .Pointer
    is unset and the mapping address is unmapped (INVALIDATED).
So that we can deploy the new APIs to Linux with minimal effort by just
invoking acpi_get_table() in acpi_get_table_with_size() and invoking
acpi_put_table() in early_acpi_os_unmap_memory(). Lv Zheng.

Link: https://github.com/acpica/acpica/commit/cac67909Signed-off-by: NLv Zheng <lv.zheng@intel.com>
Signed-off-by: NBob Moore <robert.moore@intel.com>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>
      174cc718
    • J
      dt: bindings: net: use boolean dt properties for eee broken modes · 308d3165
      由 jbrunet 提交于 
      The patches regarding eee-broken-modes was merged before all people
involved could find an agreement on the best way to move forward.

While we agreed on having a DT property to mark particular modes as broken,
the value used for eee-broken-modes mapped the phy register in very direct
way. Because of this, the concern is that it could be used to implement
configuration policies instead of describing a broken HW.

In the end, having a boolean property for each mode seems to be preferred
over one bit field value mapping the register (too) directly.

Cc: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: NJerome Brunet <jbrunet@baylibre.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      308d3165
    • J
      ratelimit: fix WARN_ON_RATELIMIT return value · 1b011e2f
      由 Jiri Slaby 提交于 
      The macro is to be used similarly as WARN_ON as:

  if (WARN_ON_RATELIMIT(condition, state))
	do_something();

One would expect only 'condition' to affect the 'if', but
WARN_ON_RATELIMIT does internally only:

  WARN_ON((condition) && __ratelimit(state))

So the 'if' is affected by the ratelimiting state too.  Fix this by
returning 'condition' in any case.

Note that nobody uses WARN_ON_RATELIMIT yet, so there is nothing to
worry about.  But I was about to use it and was a bit surprised.

Link: http://lkml.kernel.org/r/20161215093224.23126-1-jslaby@suse.czSigned-off-by: NJiri Slaby <jslaby@suse.cz>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
      1b011e2f
    • M
      ima: on soft reboot, save the measurement list · 7b8589cc
      由 Mimi Zohar 提交于 
      The TPM PCRs are only reset on a hard reboot.  In order to validate a
TPM's quote after a soft reboot (eg.  kexec -e), the IMA measurement
list of the running kernel must be saved and restored on boot.

This patch uses the kexec buffer passing mechanism to pass the
serialized IMA binary_runtime_measurements to the next kernel.

Link: http://lkml.kernel.org/r/1480554346-29071-7-git-send-email-zohar@linux.vnet.ibm.comSigned-off-by: NThiago Jung Bauermann <bauerman@linux.vnet.ibm.com>
Signed-off-by: NMimi Zohar <zohar@linux.vnet.ibm.com>
Acked-by: N"Eric W. Biederman" <ebiederm@xmission.com>
Acked-by: NDmitry Kasatkin <dmitry.kasatkin@gmail.com>
Cc: Andreas Steffen <andreas.steffen@strongswan.org>
Cc: Josh Sklar <sklar@linux.vnet.ibm.com>
Cc: Dave Young <dyoung@redhat.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Baoquan He <bhe@redhat.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Stewart Smith <stewart@linux.vnet.ibm.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
      7b8589cc
  21. 20 12月, 2016 2 次提交
  23. 19 12月, 2016 1 次提交
  25. 18 12月, 2016 7 次提交