On 08/30/2013 04:26 PM, Al Viro wrote:
On Fri, Aug 30, 2013 at 03:52:49PM -0400, Waiman Long wrote:
So it is the mnput_no_expire() function that is doing all the
lg_global_lock() calls.
Interesting... So you are getting a lot of mntput() with ->mnt_ns being
NULL? I wonder which type it
On 08/30/2013 04:54 PM, Al Viro wrote:
On Fri, Aug 30, 2013 at 01:43:11PM -0700, Linus Torvalds wrote:
On Fri, Aug 30, 2013 at 1:15 PM, Waiman Long wrote:
The prepend_path() isn't all due to getcwd. The correct profile should be
Ugh. I really think that prepend_path() should ju
On 08/30/2013 05:30 PM, Al Viro wrote:
On Fri, Aug 30, 2013 at 05:10:45PM -0400, Waiman Long wrote:
See above. You are right, but if Linus wants to turn that sucker into
reader (which is possible - see e.g. cifs build_path_from_dentry() and
its ilk), d_move() races will start to play.
Thank
On 08/30/2013 04:48 PM, Al Viro wrote:
On Fri, Aug 30, 2013 at 04:35:49PM -0400, Waiman Long wrote:
The AIM7 test was run on a set of 16 ramdisk formated with ext3
filesystem with the following mount options:
barrier=0,async,noatime,nodiratime. Maybe that is a factor.
I would be really
On 08/30/2013 10:42 PM, Al Viro wrote:
On Sat, Aug 31, 2013 at 03:35:16AM +0100, Al Viro wrote:
Aha... OK, I see what's going on. We end up with shm_mnt *not* marked
as long-living vfsmount, even though it lives forever. See if the
following helps; if it does (and I very much expect it to),
On 09/03/2013 02:01 AM, Ingo Molnar wrote:
* Waiman Long wrote:
Yes, that patch worked. It eliminated the lglock as a bottleneck in
the AIM7 workload. The lg_global_lock did not show up in the perf
profile, whereas the lg_local_lock was only 0.07%.
Just curious: what's the worst bottl
On 09/03/2013 03:09 PM, Linus Torvalds wrote:
On Tue, Sep 3, 2013 at 8:34 AM, Linus Torvalds
wrote:
I suspect the tty_ldisc_lock() could be made to go away if we care.
Heh. I just pulled the tty patches from Greg, and the locking has
changed completely.
It may actually fix your AIM7 test-cas
syms] [k] _raw_spin_lock_irqsave
|--0.94%-- rwsem_wake
1.00% reaim [kernel.kallsyms] [k] _raw_spin_lock_irq
|--88.70%-- rwsem_down_write_failed
Signed-off-by: Waiman Long
---
include/linux/rwsem.h |2 ++
lib/rwsem.c | 19 +++
2 files c
On 09/27/2013 03:32 PM, Peter Hurley wrote:
On 09/27/2013 03:00 PM, Waiman Long wrote:
With the 3.12-rc2 kernel, there is sizable spinlock contention on
the rwsem wakeup code path when running AIM7's high_systime workload
on a 8-socket 80-core DL980 (HT off) as reported by perf:
On 09/27/2013 02:09 PM, Tim Chen wrote:
On Fri, 2013-09-27 at 08:29 -0700, Paul E. McKenney wrote:
On Wed, Sep 25, 2013 at 03:10:49PM -0700, Tim Chen wrote:
We will need the MCS lock code for doing optimistic spinning for rwsem.
Extracting the MCS code from mutex.c and put into its own file all
On 09/28/2013 12:34 AM, Jason Low wrote:
Also, below is what the mcs_spin_lock() and mcs_spin_unlock()
functions would look like after applying the proposed changes.
static noinline
void mcs_spin_lock(struct mcs_spin_node **lock, struct mcs_spin_node *node)
{
struct mcs_spin_node *prev;
On 09/28/2013 03:33 PM, Linus Torvalds wrote:
On Sat, Sep 28, 2013 at 12:21 PM, Ingo Molnar wrote:
If we do that then I suspect the next step will be queued rwlocks :-/ The
current rwlock_t implementation is rather primitive by modern standards.
(We'd probably have killed rwlock_t long ago if n
On 09/30/2013 03:05 AM, Ingo Molnar wrote:
* Michel Lespinasse wrote:
That said, I am very scared of using rwlock_t here, and I would much
prefer we choose a fair lock (either spinlock or a new rwlock
implementation which guarantees not to starve any locker thread)
Given how few users rwlock_
|--88.70%-- rwsem_down_write_failed
Signed-off-by: Waiman Long
---
include/linux/rwsem.h |2 ++
lib/rwsem.c | 29 +
2 files changed, 31 insertions(+), 0 deletions(-)
diff --git a/include/linux/rwsem.h b/include/linux/rwsem.h
index 0616ffe..e25792e 100644
--- a/inc
On 09/30/2013 12:10 PM, Jason Low wrote:
On Mon, 2013-09-30 at 11:51 -0400, Waiman Long wrote:
On 09/28/2013 12:34 AM, Jason Low wrote:
Also, below is what the mcs_spin_lock() and mcs_spin_unlock()
functions would look like after applying the proposed changes.
static noinline
void
On 09/30/2013 03:23 PM, Tim Chen wrote:
On Mon, 2013-09-30 at 20:14 +0200, Peter Zijlstra wrote:
On Mon, Sep 30, 2013 at 10:10:27AM -0700, Tim Chen wrote:
Here's the exim workload data:
rwsem improvment:
Waimain's patch:+2.0%
Alex+Tim's patchset:+4.8%
Waiman+Alex+Tim:+5.3%
On 09/30/2013 03:29 PM, Linus Torvalds wrote:
So with all the lockref work, we now avoid the dentry d_lock for
almost all normal cases.
There is one single remaining common case, though: the final dput()
when the dentry count goes down to zero, and we need to check if we
are supposed to get rid
On 09/30/2013 03:47 PM, Tim Chen wrote:
My qrwlock doesn't enable qrwlock by default. You have to use menuconfig
to explicitly enable it. Have you done that when you build the test
kernel? I am thinking of explicitly enabling it for x86 if the anon-vma
lock is converted back to a rwlock.
Yes, I
those
functions.
Signed-off-by: Waiman Long
---
init/Kconfig | 13 +++
kernel/sched/core.c |1 +
kernel/sched/fair.c | 55 +-
kernel/sched/sched.h | 23
4 files changed, 91 insertions(+), 1 deletions(-)
s (from 139.8s to
139s) in this case.
Signed-off-by: Waiman Long
---
kernel/sched/fair.c | 29 ++---
kernel/sched/sched.h |6 --
2 files changed, 26 insertions(+), 9 deletions(-)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 68f1609..979f04d 10
These 2 patches try to reduce contention on the load_avg and
runnable_avg values of the task_group structure in large NUMA
machine by:
patch 1: Move them into their own cacheline & reduce actual read/write
to them.
patch 2: Reduce the frequecy of doing the average computation.
Wa
probably due to the spinner locking and queuing overhead.
Signed-off-by: Waiman Long
Acked-by: Rik van Riel
---
include/linux/mutex.h |3 ++
include/linux/sched.h |3 ++
kernel/mutex.c| 78 -
kernel/sched/core.c | 24 +
to kernel/mutex.c
which is where they should belong.
Signed-off-by: Waiman Long
---
include/linux/sched.h |4 ---
kernel/mutex.c | 62 ++
kernel/sched/core.c | 63 ---
kernel/sched
atch 5 removes SCHED_FEAT_OWNER_SPIN which was just an earlier hack
for testing purpose. It also moves the mutex spinning code back to
mutex.c.
Waiman Long (5):
mutex: Make more scalable by doing less atomic operations
mutex: Queue mutex spinners with MCS lock to reduce cacheline
contention
m
In response to the review comment from Davidlohr, this patch will
remove the new typedefs introduced by patch 2. It also removes an
unnecessary barrier() call.
Signed-off-by: Waiman Long
---
kernel/mutex.c | 25 +++--
1 files changed, 11 insertions(+), 14 deletions
this is not the case, this patch should be dropped.
Signed-off-by: Waiman Long
---
arch/x86/include/asm/mutex.h | 10 --
kernel/mutex.c |9 ++---
2 files changed, 2 insertions(+), 17 deletions(-)
diff --git a/arch/x86/include/asm/mutex.h b/arch/x86/include/asm
% | +4.2% |
+--+---++-+
Signed-off-by: Waiman Long
Reviewed-by: Davidlohr Bueso
Reviewed-by: Rik van Riel
---
arch/x86/include/asm/mutex.h | 10 ++
kernel/mutex.c | 19 ---
2 files changed, 26
ance for the same workloads.
Patch 4 is an optional one for backing out architecture specific
check in patch 2, if so desired.
Waiman Long (4):
mutex: Move mutex spinning code from sched/core.c back to mutex.c
mutex: Make more scalable by doing less atomic operations
mutex: Queue mutex spinn
% | +4.2% |
+--+---++-+
Signed-off-by: Waiman Long
Reviewed-by: Davidlohr Bueso
Reviewed-by: Rik van Riel
---
arch/x86/include/asm/mutex.h | 10 ++
kernel/mutex.c | 19 ---
2 files changed, 26
this is not the case, this patch should be dropped.
Signed-off-by: Waiman Long
---
arch/x86/include/asm/mutex.h | 10 --
kernel/mutex.c |9 ++---
2 files changed, 2 insertions(+), 17 deletions(-)
diff --git a/arch/x86/include/asm/mutex.h b/arch/x86/include/asm
probably due to the spinner locking and queuing overhead.
Signed-off-by: Waiman Long
Acked-by: Rik van Riel
---
include/linux/mutex.h |3 ++
kernel/mutex.c| 91 -
2 files changed, 93 insertions(+), 1 deletions(-)
diff --git a/include/linu
to kernel/mutex.c
which is where they should belong.
Signed-off-by: Waiman Long
---
include/linux/sched.h |1 -
kernel/mutex.c | 46 ++
kernel/sched/core.c | 45 -
kernel/sched/features.h
On 04/17/2013 03:18 AM, Ingo Molnar wrote:
* Waiman Long wrote:
As mentioned by Ingo, the SCHED_FEAT_OWNER_SPIN scheduler feature
bit was really just an early hack to make with/without mutex-spinning
testable. So it is no longer necessary.
This patch removes the SCHED_FEAT_OWNER_SPIN feature
-2.8% | -3.3% | -2.1% |
| shared | -0.6% | -0.2% | +0.2% |
+--+---++-+
There are slight drops in performance for fsever and new_fserver
workloads, but slight increase in the high_sys
On 06/14/2013 11:37 AM, Linus Torvalds wrote:
On Fri, Jun 14, 2013 at 8:00 AM, Waiman Long wrote:
On 06/12/2013 08:59 PM, Linus Torvalds wrote:
Ho humm.. interesting. I was talking about wanting to mix atomics and
spinlocks earlier in this thread due to space constraints, and it
strikes me
On 06/14/2013 09:26 PM, Benjamin Herrenschmidt wrote:
On Fri, 2013-06-14 at 14:17 -0400, Waiman Long wrote:
With some minor changes, the current patch can be modified to support
debugging lock for 32-bit system. For 64-bit system, we can apply a
similar concept for debugging lock with
On 05/03/2013 10:07 AM, Waiman Long wrote:
On 04/10/2013 02:26 PM, Waiman Long wrote:
While running the high_systime workload of the AIM7 benchmark on
a 2-socket 12-core Westmere x86-64 machine running 3.8.2 kernel,
it was found that a pretty sizable amount of time was spent in the
SELinux code
On 06/05/2013 10:59 AM, Stephen Smalley wrote:
Can you take the core logic into a helper function within ebitmap.c?
Otherwise you are directly exposing ebitmap internals to the mls code.
Sure. I will move the logic to ebitmap.c & send out an updated patch.
Thank for the quick response.
Rega
00-2000 users |
+--+---++-+
| high_systime | +0.1% | +0.9% | +2.6% |
+--+---++-+
Signed-off-by: Waiman Long
---
security/selinux/ss/ebitmap.c | 35 ++-
security/selinux/ss/ebitmap.h |2
overhead ratio at 1/4. This may
also improve performance a little bit by making node to node traversal
less frequent (< 2) as more bits are available in each node.
Signed-off-by: Waiman Long
---
security/selinux/ss/ebitmap.h |8 +++-
1 files changed, 7 insertions(+), 1 deletions(-)
d
sers | 1100-2000 users |
+--+---++-+
| high_systime | +0.1% | +0.9% | +2.6% |
+--+---++-+
Signed-off-by: Waiman Long
---
security/selinux/ss/eb
The d_lock and d_count fields of the struct dentry in dcache.h was
modified to use the new mechanism. This serves as an example of how
to convert existing spinlock and reference count combo to use the
new way of locklessly updating the reference count.
Signed-off-by: Waiman Long
Waiman Lo
UG_SPINLOCK=n
4. x86 32-bit SMP, CONFIG_DEBUG_SPINLOCK=n
5. x86 32-bit SMP, CONFIG_DEBUG_SPINLOCK=y
Signed-off-by: Waiman Long
---
include/linux/spinlock_refcount.h | 145 +
include/linux/spinlock_types.h| 19 +
2 files changed, 164 insertions(
.7% | -0.9% |
| shared | -0.1% | 0.0% | -0.1% |
+--+---++-+
There are slight drops in performance for fsever and new_fserver
workloads, but slight increase in the high_systime workload.
Signed-
On 05/26/2013 10:09 PM, Dave Chinner wrote:
On Thu, May 23, 2013 at 05:34:23PM -0400, Waiman Long wrote:
On 05/23/2013 05:42 AM, Dave Chinner wrote:
What was it I said about this patchset when you posted it to speed
up an Oracle benchmark back in february? I'll repeat:
"Nobody
On 05/29/2013 12:13 PM, Andi Kleen wrote:
The d_path() is called by perf_event_mmap_event() which translates
VMA to its file path for memory segments backed by files. As perf is
not just for sampling data within the kernel, it can also be used
for checking access pattern in the user space. As a r
On 05/29/2013 12:18 PM, Simo Sorce wrote:
On Wed, 2013-05-29 at 11:55 -0400, Waiman Long wrote:
My patch set consists of 2 different changes. The first one is to avoid
taking the d_lock lock when updating the reference count in the
dentries. This particular change also benefit some other
On 05/29/2013 12:56 PM, Andi Kleen wrote:
On Wed, May 29, 2013 at 12:18:09PM -0400, Simo Sorce wrote:
To be honest a workload base on /etc/passwd or /etc/group is completely
artificial, in actual usage, if you really have such access you use
nscd or sssd with their shared memory caches to compl
On 05/29/2013 02:46 PM, J. Bruce Fields wrote:
On Wed, May 29, 2013 at 11:55:14AM -0400, Waiman Long wrote:
On 05/26/2013 10:09 PM, Dave Chinner wrote:
On Thu, May 23, 2013 at 05:34:23PM -0400, Waiman Long wrote:
On 05/23/2013 05:42 AM, Dave Chinner wrote:
What was it I said about this
On 05/29/2013 05:19 PM, Jörn Engel wrote:
On Wed, 29 May 2013 22:37:00 +0200, Andi Kleen wrote:
As Dave said before, is the last path component sufficient? Or how
about an inode number?
Neither works, the profiler needs to find the file and read it.
Ignoring all the complexity this would caus
On 06/19/2013 12:50 PM, Waiman Long wrote:
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 updat
On 06/07/2013 02:09 PM, Stephen Smalley wrote:
Can't you unify this logic with the nearly identical logic below?
By separating the logic, we can usually get the best performance out of
the two separate loops. If code size and maintainability are a bigger
concern, I can certainly merge the l
.9% | +2.6% |
+--+---++-+
Signed-off-by: Waiman Long
---
security/selinux/ss/ebitmap.c | 20 ++--
security/selinux/ss/ebitmap.h |2 +-
security/selinux/ss/mls.c | 22 +++---
security/selinux/ss/mls_types.h |2 +-
4 files changed, 27 insertio
+0.9% | +2.6% |
+--+---++-+
Signed-off-by: Waiman Long
---
security/selinux/ss/ebitmap.c | 20 ++--
security/selinux/ss/ebitmap.h |2 +-
security/selinux/ss/mls.c | 22 +++---
security/selinux/ss/ml
On 06/09/2013 03:36 PM, Paul E. McKenney wrote:
Breaking up locks is better than implementing high-contention locks, but
if we must have high-contention locks, why not make them automatically
switch between light-weight ticket locks at low contention and queued
locks at high contention?
This com
On 06/11/2013 12:20 PM, Steven Rostedt wrote:
diff --git a/arch/x86/include/asm/spinlock_types.h
b/arch/x86/include/asm/spinlock_types.h
index ad0ad07..cdaefdd 100644
--- a/arch/x86/include/asm/spinlock_types.h
+++ b/arch/x86/include/asm/spinlock_types.h
@@ -7,12 +7,18 @@
#include
-#if (CON
On 06/11/2013 12:36 PM, Paul E. McKenney wrote:
I am a bit concern about the size of the head queue table itself.
RHEL6, for example, had defined CONFIG_NR_CPUS to be 4096 which mean
a table size of 256. Maybe it is better to dynamically allocate the
table at init time depending on the actual n
ng
the update. It is because there will still be some fair amount of
contention with only one attempt.
Signed-off-by: Waiman Long
---
include/asm-generic/spinlock_refcount.h | 98 +
include/linux/spinlock_refcount.h | 159 +
kernel/Kconfig.locks
-by: Waiman Long
---
arch/x86/Kconfig |3 +++
arch/x86/include/asm/spinlock_refcount.h |1 +
2 files changed, 4 insertions(+), 0 deletions(-)
create mode 100644 arch/x86/include/asm/spinlock_refcount.h
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index fe120da
c and everything should just work.
Signed-off-by: Waiman Long
---
arch/powerpc/platforms/cell/spufs/inode.c |6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/arch/powerpc/platforms/cell/spufs/inode.c
b/arch/powerpc/platforms/cell/spufs/inode.c
index 35f77a4..3597c4b 100644
t are replaced by the d_ret_count() calls.
There is no change in logic and everything should just work.
Signed-off-by: Waiman Long
---
fs/dcookies.c |8
fs/fs-writeback.c |4 ++--
fs/libfs.c| 36 ++--
fs/namei.c|
t are replaced by the d_ret_count() calls.
There is no change in logic and everything should just work.
Signed-off-by: Waiman Long
---
fs/coda/cache.c |4 ++--
fs/coda/dir.c |4 ++--
2 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/fs/coda/cache.c b/fs/coda/cache.c
c and everything should just work.
Signed-off-by: Waiman Long
---
fs/9p/fid.c | 14 +++---
1 files changed, 7 insertions(+), 7 deletions(-)
diff --git a/fs/9p/fid.c b/fs/9p/fid.c
index d51ec9f..fc23dd3 100644
--- a/fs/9p/fid.c
+++ b/fs/9p/fid.c
@@ -48,9 +48,9 @@ static inline void __add_fid(s
t are replaced by the d_ret_count() calls.
There is no change in logic and everything should just work.
Signed-off-by: Waiman Long
---
fs/ceph/caps.c |8
fs/ceph/debugfs.c|8
fs/ceph/dir.c| 34 +-
fs/ceph/export.c |
| +1.2% | +0.3% |
| shared | +0.8% | -0.3% | 0.0% |
+--+---++-+
There are slight drops in performance for the five_sec workload,
but slight increase in the high_systime workload.
Signed-off-
c and everything should just work.
Signed-off-by: Waiman Long
---
fs/afs/dir.c |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/fs/afs/dir.c b/fs/afs/dir.c
index 7a465ed..fa235e0 100644
--- a/fs/afs/dir.c
+++ b/fs/afs/dir.c
@@ -705,9 +705,9 @@ out_skip:
/* the diren
c and everything should just work.
Signed-off-by: Waiman Long
---
drivers/infiniband/hw/ipath/ipath_fs.c |6 +++---
drivers/infiniband/hw/qib/qib_fs.c |6 +++---
2 files changed, 6 insertions(+), 6 deletions(-)
diff --git a/drivers/infiniband/hw/ipath/ipath_fs.c
b/drivers/infiniband/hw/
c and everything should just work.
Signed-off-by: Waiman Long
---
fs/cifs/dir.c | 10 +-
1 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/fs/cifs/dir.c b/fs/cifs/dir.c
index 5699b50..51d9f7e 100644
--- a/fs/cifs/dir.c
+++ b/fs/cifs/dir.c
@@ -115,10 +115,10 @@ cifs_bp_rename_
t are replaced by the d_ret_count() calls.
There is no change in logic and everything should just work.
Signed-off-by: Waiman Long
---
fs/autofs4/autofs_i.h | 24
fs/autofs4/expire.c | 48
fs/autofs4/root.c |
Because of the changes made in dcache.h header file, files that use the
d_count field of the dentry structure need to be changed accordingly.
References to d_count are replaced by the d_ret_count() calls.
There is no change in logic and everything should just work.
Signed-off-by: Waiman Long
c and everything should just work.
Signed-off-by: Waiman Long
---
security/selinux/selinuxfs.c | 14 +++---
1 files changed, 7 insertions(+), 7 deletions(-)
diff --git a/security/selinux/selinuxfs.c b/security/selinux/selinuxfs.c
index ff42773..c83a201 100644
--- a/security/selinux/selinu
t are replaced by the d_ret_count() calls.
There is no change in logic and everything should just work.
Signed-off-by: Waiman Long
---
fs/nfs/dir.c | 14 +++---
fs/nfs/getroot.c |8
fs/nfs/namespace.c | 16
fs/nfs/unlink.c|
c and everything should just work.
Signed-off-by: Waiman Long
---
fs/exportfs/expfs.c |8
1 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/fs/exportfs/expfs.c b/fs/exportfs/expfs.c
index 262fc99..9cf6a44 100644
--- a/fs/exportfs/expfs.c
+++ b/fs/exportfs/expfs.c
@@ -114,15 +1
Because of the changes made in dcache.h header file, files that use the
d_count field of the dentry structure need to be changed accordingly.
References to d_count are replaced by the d_ret_count() calls.
There is no change in logic and everything should just work.
Signed-off-by: Waiman Long
c and everything should just work.
Signed-off-by: Waiman Long
---
net/sunrpc/rpc_pipe.c |6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/net/sunrpc/rpc_pipe.c b/net/sunrpc/rpc_pipe.c
index e7ce4b3..58bc128 100644
--- a/net/sunrpc/rpc_pipe.c
+++ b/net/sunrpc/rpc_pipe.c
@@ -4
c and everything should just work.
Signed-off-by: Waiman Long
---
fs/notify/fsnotify.c |8
fs/notify/vfsmount_mark.c| 24
include/linux/fsnotify_backend.h |6 +++---
3 files changed, 19 insertions(+), 19 deletions(-)
diff --git a/fs/n
t are replaced by the d_ret_count() calls.
There is no change in logic and everything should just work.
Signed-off-by: Waiman Long
---
fs/locks.c |2 +-
include/linux/fs.h |4 ++--
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/fs/locks.c b/fs/locks.c
index cb
c and everything should just work.
Signed-off-by: Waiman Long
---
kernel/cgroup.c | 10 +-
1 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/kernel/cgroup.c b/kernel/cgroup.c
index a7c9e6d..8935c05 100644
--- a/kernel/cgroup.c
+++ b/kernel/cgroup.c
@@ -974,11 +974,11 @@ static
t are replaced by the d_ret_count() calls.
There is no change in logic and everything should just work.
Signed-off-by: Waiman Long
---
fs/configfs/configfs_internal.h |4 ++--
fs/configfs/dir.c |2 +-
fs/configfs/inode.c |6 +++---
3 files changed, 6 inser
c and everything should just work.
Signed-off-by: Waiman Long
---
fs/ocfs2/dcache.c |6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/fs/ocfs2/dcache.c b/fs/ocfs2/dcache.c
index ef99972..4f72906 100644
--- a/fs/ocfs2/dcache.c
+++ b/fs/ocfs2/dcache.c
@@ -173,17 +173,17 @@ s
c and everything should just work.
Signed-off-by: Waiman Long
---
fs/fat/namei_vfat.c |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/fs/fat/namei_vfat.c b/fs/fat/namei_vfat.c
index 2da9520..722ca30 100644
--- a/fs/fat/namei_vfat.c
+++ b/fs/fat/namei_vfat.c
@@ -34,10 +
c and everything should just work.
Signed-off-by: Waiman Long
---
fs/ncpfs/dir.c |6 +++---
fs/ncpfs/ncplib_kernel.h |8
2 files changed, 7 insertions(+), 7 deletions(-)
diff --git a/fs/ncpfs/dir.c b/fs/ncpfs/dir.c
index 6792ce1..c78b801 100644
--- a/fs/ncpfs/dir.c
+++ b/fs/
entry structure.
Thank to Thomas Gleixner, Andi Kleen and Linus for their valuable
input in shaping this patchset.
Signed-off-by: Waiman Long
Waiman Long (25):
spinlock: A new lockref structure for lockless update of refcount
spinlock: Enable x86 architecture to do lockless refcount update
On 07/03/2013 04:37 PM, Linus Torvalds wrote:
This patch grew a lot, and that seems to be mainly because of bad reasons.
That is the main reason why I choose to implement it the way it was in
my previous version. As I add one more level to access d_lock and
d_count, I need to change a lot mor
n the short
workload of the AIM7 benchmark running on a 8-socket 80-core DL980
machine on a 3.10-based kernel. There were still a few percentage
points of contention in d_path() and getcwd syscall left due to their
use of the rename_lock.
Signed-off-by: Waiman Long
Waiman Long (4):
seqlock: Ad
w
blocking reader will not block other non-blocking readers, but will
block other blocking readers and writers.
Signed-off-by: Waiman Long
---
include/linux/seqlock.h | 65 +++---
1 files changed, 60 insertions(+), 5 deletions(-)
diff --git a/include/linux/s
tool is used to record performance data on
a large SMP system with many cores, converting rename_lock to use
read/write lock will eliminate this bottleneck and the skewing of
perf trace data.
Signed-off-by: Waiman Long
---
fs/dcache.c |2 +-
1 files changed, 1 insertions(+), 1 deletions
. d_path
4. dentry_path_raw
5. dentry_path
6. getcwd syscall
Signed-off-by: Waiman Long
---
fs/dcache.c | 26 +-
1 files changed, 13 insertions(+), 13 deletions(-)
diff --git a/fs/dcache.c b/fs/dcache.c
index f09b908..480c81f 100644
--- a/fs/dcache.c
+++ b/fs/dcache.c
-off-by: Waiman Long
---
include/linux/seqlock.h | 118 --
1 files changed, 92 insertions(+), 26 deletions(-)
diff --git a/include/linux/seqlock.h b/include/linux/seqlock.h
index 26be0d9..a1fd45c 100644
--- a/include/linux/seqlock.h
+++ b/include/linux
ructure to use the lockref
structure to improve performance for high contention
cases
Thank to Thomas Gleixner, Andi Kleen and Linus for their valuable
input in shaping this patchset.
Signed-off-by: Waiman Long
Waiman Long (12):
spinlock: A new lockref structure f
Because of the d_count name change made in dcache.h, all references
to d_count have to be changed to d_refcount. There is no change in
logic and everything should just work.
Signed-off-by: Waiman Long
---
fs/autofs4/expire.c |8
fs/autofs4/root.c |2 +-
2 files changed, 5
d_count field name to a different
one which is unique in the kernel source code. The new name will
be d_refcount.
Signed-off-by: Waiman Long
---
fs/dcache.c| 54
fs/namei.c |6 ++--
include/linux/dcache.h |6 ++--
3
Because of the d_count name change made in dcache.h, all references
to d_count have to be changed to d_refcount. There is no change in
logic and everything should just work.
Signed-off-by: Waiman Long
---
fs/coda/dir.c |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a
Because of the d_count name change made in dcache.h, all references
to d_count have to be changed to d_refcount. There is no change in
logic and everything should just work.
Signed-off-by: Waiman Long
---
fs/nilfs2/super.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git
Because of the d_count name change made in dcache.h, all references
to d_count have to be changed to d_refcount. There is no change in
logic and everything should just work.
Signed-off-by: Waiman Long
---
fs/configfs/dir.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git
Because of the d_count name change made in dcache.h, all references
to d_count have to be changed to d_refcount. There is no change in
logic and everything should just work.
Signed-off-by: Waiman Long
---
fs/ecryptfs/inode.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff
Because of the d_count name change made in dcache.h, all references
to d_count have to be changed to d_refcount. There is no change in
logic and everything should just work.
Signed-off-by: Waiman Long
---
fs/locks.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/fs
Because of the d_count name change made in dcache.h, all references
to d_count have to be changed to d_refcount. There is no change in
logic and everything should just work.
Signed-off-by: Waiman Long
---
fs/nfs/dir.c|8
fs/nfs/unlink.c |2 +-
2 files changed, 5 insertions
Because of the d_count name change made in dcache.h, all references
to d_count have to be changed to d_refcount. There is no change in
logic and everything should just work.
Signed-off-by: Waiman Long
---
fs/ceph/inode.c |4 ++--
fs/ceph/mds_client.c |2 +-
2 files changed, 3
won't work for those macros and so the errors should be ignored.
Signed-off-by: Waiman Long
---
fs/dcache.c| 18 --
include/linux/dcache.h | 17 ++---
2 files changed, 22 insertions(+), 13 deletions(-)
diff --git a/fs/dcache.c b/fs/dcache.c
index 61d
101 - 200 of 3826 matches
Mail list logo
