On 08/05/2013 11:12 PM, Waiman Long wrote:
v6->v7:
- Substantially reduce the number of patches from 14 to 4 because a
lot of the minor filesystem related changes had been merged to
v3.11-rc1.
- Remove architecture specific customization (LOCKREF_WAIT_SHIFT&
LOCKREF_RETRY_COUNT).
- Tune single-thread performance of lockref_put/get to within 10%
of old lock->update->unlock code.
v5->v6:
- Add a new GENERIC_SPINLOCK_REFCOUNT config parameter for using the
generic implementation.
- Add two parameters LOCKREF_WAIT_SHIFT and LOCKREF_RETRY_COUNT which
can be specified differently for each architecture.
- Update various spinlock_refcount.* files to incorporate review
comments.
- Replace reference of d_refcount() macro in Lustre filesystem code in
the staging tree to use the new d_count() helper function.
v4->v5:
- Add a d_count() helper for readonly access of reference count and
change all references to d_count outside of dcache.c, dcache.h
and namei.c to use d_count().
v3->v4:
- Replace helper function access to d_lock and d_count by using
macros to redefine the old d_lock name to the spinlock and new
d_refcount name to the reference count. This greatly reduces the
size of this patchset from 25 to 12 and make it easier to review.
v2->v3:
- Completely revamp the packaging by adding a new lockref data
structure that combines the spinlock with the reference
count. Helper functions are also added to manipulate the new data
structure. That results in modifying over 50 files, but the changes
were trivial in most of them.
- Change initial spinlock wait to use a timeout.
- Force 64-bit alignment of the spinlock& reference count structure.
- Add a new way to use the combo by using a new union and helper
functions.
v1->v2:
- Add one more layer of indirection to LOCK_WITH_REFCOUNT macro.
- Add __LINUX_SPINLOCK_REFCOUNT_H protection to spinlock_refcount.h.
- Add some generic get/put macros into spinlock_refcount.h.
This patchset supports a generic mechanism to atomically update
a reference count that is protected by a spinlock without actually
acquiring the lock itself. If the update doesn't succeeed, the caller
will have to acquire the lock and update the reference count in the
the old way. This will help in situation where there is a lot of
spinlock contention because of frequent reference count update.
The d_lock and d_count fields of the struct dentry in dcache.h was
modified to use the new lockref data structure and the d_lock name
is now a macro to the actual spinlock.
This patch set causes significant performance improvement in the
short workload of the AIM7 benchmark on a 8-socket x86-64 machine
with 80 cores.
Thank to Thomas Gleixner, Andi Kleen and Linus for their valuable
input in shaping this patchset.
Signed-off-by: Waiman Long<Waiman.Long@xxxxxx>
Waiman Long (4):
spinlock: A new lockref structure for lockless update of refcount
spinlock: Enable x86 architecture to do lockless refcount update
dcache: replace d_lock/d_count by d_lockcnt
dcache: Enable lockless update of dentry's refcount
arch/x86/Kconfig | 3 +
fs/dcache.c | 78 +++++++------
fs/namei.c | 6 +-
include/asm-generic/spinlock_refcount.h | 46 +++++++
include/linux/dcache.h | 22 ++--
include/linux/spinlock_refcount.h | 126 ++++++++++++++++++++
kernel/Kconfig.locks | 15 +++
lib/Makefile | 2 +
lib/spinlock_refcount.c | 198 +++++++++++++++++++++++++++++++
9 files changed, 449 insertions(+), 47 deletions(-)
create mode 100644 include/asm-generic/spinlock_refcount.h
create mode 100644 include/linux/spinlock_refcount.h
create mode 100644 lib/spinlock_refcount.c
So far, I haven't heard back anything about if further change and
improvement is needed for this patch set. Any comment or feedback will
be appreciated.
Thank in advance for your time.
Regards,
Longman
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
