While parallel applications tend to align their data on the cache
boundary, they tend not to align on the page or THP boundary.
Consequently tasks that partition their data can still "false-share"
pages presenting a problem for optimal NUMA placement.
This patch uses NUMA hinting faults to chain tasks together into
numa_groups. As well as storing the NID a task was running on when
accessing a page a truncated representation of the faulting PID is
stored. If subsequent faults are from different PIDs it is reasonable
to assume that those two tasks share a page and are candidates for
being grouped together. Note that this patch makes no scheduling
decisions based on the grouping information.
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Link: http://lkml.kernel.org/r/1381141781-10992-44-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Change the per page last fault tracking to use cpu,pid instead of
nid,pid. This will allow us to try and lookup the alternate task more
easily. Note that even though it is the cpu that is store in the page
flags that the mpol_misplaced decision is still based on the node.
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Link: http://lkml.kernel.org/r/1381141781-10992-43-git-send-email-mgorman@suse.de
[ Fixed build failure on 32-bit systems. ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The load balancer will spread workloads across multiple NUMA nodes,
in order to balance the load on the system. This means that sometimes
a task's preferred node has available capacity, but moving the task
there will not succeed, because that would create too large an imbalance.
In that case, other NUMA nodes need to be considered.
Signed-off-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Mel Gorman <mgorman@suse.de>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-42-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
A tasks preferred node is selected based on the number of faults
recorded for a node but the actual task_numa_migate() conducts a global
search regardless of the preferred nid. This patch checks if the
preferred nid has capacity and if so, searches for a CPU within that
node. This avoids a global search when the preferred node is not
overloaded.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-41-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This patch implements a system-wide search for swap/migration candidates
based on total NUMA hinting faults. It has a balance limit, however it
doesn't properly consider total node balance.
In the old scheme a task selected a preferred node based on the highest
number of private faults recorded on the node. In this scheme, the preferred
node is based on the total number of faults. If the preferred node for a
task changes then task_numa_migrate will search the whole system looking
for tasks to swap with that would improve both the overall compute
balance and minimise the expected number of remote NUMA hinting faults.
Not there is no guarantee that the node the source task is placed
on by task_numa_migrate() has any relationship to the newly selected
task->numa_preferred_nid due to compute overloading.
Signed-off-by: Mel Gorman <mgorman@suse.de>
[ Do not swap with tasks that cannot run on source cpu]
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
[ Fixed compiler warning on UP. ]
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-40-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Use the new stop_two_cpus() to implement migrate_swap(), a function that
flips two tasks between their respective cpus.
I'm fairly sure there's a less crude way than employing the stop_two_cpus()
method, but everything I tried either got horribly fragile and/or complex. So
keep it simple for now.
The notable detail is how we 'migrate' tasks that aren't runnable
anymore. We'll make it appear like we migrated them before they went to
sleep. The sole difference is the previous cpu in the wakeup path, so we
override this.
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Mel Gorman <mgorman@suse.de>
Link: http://lkml.kernel.org/r/1381141781-10992-39-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
NUMA hinting faults will not migrate a shared executable page mapped by
multiple processes on the grounds that the data is probably in the CPU
cache already and the page may just bounce between tasks running on multipl
nodes. Even if the migration is avoided, there is still the overhead of
trapping the fault, updating the statistics, making scheduler placement
decisions based on the information etc. If we are never going to migrate
the page, it is overhead for no gain and worse a process may be placed on
a sub-optimal node for shared executable pages. This patch avoids trapping
faults for shared libraries entirely.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-36-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
When a task is already running on its preferred node, increment
numa_migrate_seq to indicate that the task is settled if migration is
temporarily disabled, and memory should migrate towards it.
Signed-off-by: Rik van Riel <riel@redhat.com>
[ Only increment migrate_seq if migration temporarily disabled. ]
Signed-off-by: Mel Gorman <mgorman@suse.de>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-35-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
When a preferred node is selected for a tasks there is an attempt to migrate
the task to a CPU there. This may fail in which case the task will only
migrate if the active load balancer takes action. This may never happen if
the conditions are not right. This patch will check at NUMA hinting fault
time if another attempt should be made to migrate the task. It will only
make an attempt once every five seconds.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-34-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This patch replaces find_idlest_cpu_node with task_numa_find_cpu.
find_idlest_cpu_node has two critical limitations. It does not take the
scheduling class into account when calculating the load and it is unsuitable
for using when comparing loads between NUMA nodes.
task_numa_find_cpu uses similar load calculations to wake_affine() when
selecting the least loaded CPU within a scheduling domain common to the
source and destimation nodes. It avoids causing CPU load imbalances in
the machine by refusing to migrate if the relative load on the target
CPU is higher than the source CPU.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-33-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
There is a 90% regression observed with a large Oracle performance test
on a 4 node system. Profiles indicated that the overhead was due to
contention on sp_lock when looking up shared memory policies. These
policies do not have the appropriate flags to allow them to be
automatically balanced so trapping faults on them is pointless. This
patch skips VMAs that do not have MPOL_F_MOF set.
[riel@redhat.com: Initial patch]
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reported-and-tested-by: Joe Mario <jmario@redhat.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-32-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The load balancer can move tasks between nodes and does not take NUMA
locality into account. With automatic NUMA balancing this may result in the
tasks working set being migrated to the new node. However, as the fault
buffer will still store faults from the old node the schduler may decide to
reset the preferred node and migrate the task back resulting in more
migrations.
The ideal would be that the scheduler did not migrate tasks with a heavy
memory footprint but this may result nodes being overloaded. We could
also discard the fault information on task migration but this would still
cause all the tasks working set to be migrated. This patch simply avoids
migrating the memory for a short time after a task is migrated.
Signed-off-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Mel Gorman <mgorman@suse.de>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-31-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Ideally it would be possible to distinguish between NUMA hinting faults that
are private to a task and those that are shared. If treated identically
there is a risk that shared pages bounce between nodes depending on
the order they are referenced by tasks. Ultimately what is desirable is
that task private pages remain local to the task while shared pages are
interleaved between sharing tasks running on different nodes to give good
average performance. This is further complicated by THP as even
applications that partition their data may not be partitioning on a huge
page boundary.
To start with, this patch assumes that multi-threaded or multi-process
applications partition their data and that in general the private accesses
are more important for cpu->memory locality in the general case. Also,
no new infrastructure is required to treat private pages properly but
interleaving for shared pages requires additional infrastructure.
To detect private accesses the pid of the last accessing task is required
but the storage requirements are a high. This patch borrows heavily from
Ingo Molnar's patch "numa, mm, sched: Implement last-CPU+PID hash tracking"
to encode some bits from the last accessing task in the page flags as
well as the node information. Collisions will occur but it is better than
just depending on the node information. Node information is then used to
determine if a page needs to migrate. The PID information is used to detect
private/shared accesses. The preferred NUMA node is selected based on where
the maximum number of approximately private faults were measured. Shared
faults are not taken into consideration for a few reasons.
First, if there are many tasks sharing the page then they'll all move
towards the same node. The node will be compute overloaded and then
scheduled away later only to bounce back again. Alternatively the shared
tasks would just bounce around nodes because the fault information is
effectively noise. Either way accounting for shared faults the same as
private faults can result in lower performance overall.
The second reason is based on a hypothetical workload that has a small
number of very important, heavily accessed private pages but a large shared
array. The shared array would dominate the number of faults and be selected
as a preferred node even though it's the wrong decision.
The third reason is that multiple threads in a process will race each
other to fault the shared page making the fault information unreliable.
Signed-off-by: Mel Gorman <mgorman@suse.de>
[ Fix complication error when !NUMA_BALANCING. ]
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-30-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
task_numa_work skips small VMAs. At the time the logic was to reduce the
scanning overhead which was considerable. It is a dubious hack at best.
It would make much more sense to cache where faults have been observed
and only rescan those regions during subsequent PTE scans. Remove this
hack as motivation to do it properly in the future.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-29-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
task_numa_placement checks current->mm but after buffers for faults
have already been uselessly allocated. Move the check earlier.
[peterz@infradead.org: Identified the problem]
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-27-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Ideally it would be possible to distinguish between NUMA hinting faults
that are private to a task and those that are shared. This patch prepares
infrastructure for separately accounting shared and private faults by
allocating the necessary buffers and passing in relevant information. For
now, all faults are treated as private and detection will be introduced
later.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-26-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
A preferred node is selected based on the node the most NUMA hinting
faults was incurred on. There is no guarantee that the task is running
on that node at the time so this patch rescheules the task to run on
the most idle CPU of the selected node when selected. This avoids
waiting for the balancer to make a decision.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-25-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Just as "sched: Favour moving tasks towards the preferred node" favours
moving tasks towards nodes with a higher number of recorded NUMA hinting
faults, this patch resists moving tasks towards nodes with lower faults.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-24-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This patch favours moving tasks towards NUMA node that recorded a higher
number of NUMA faults during active load balancing. Ideally this is
self-reinforcing as the longer the task runs on that node, the more faults
it should incur causing task_numa_placement to keep the task running on that
node. In reality a big weakness is that the nodes CPUs can be overloaded
and it would be more efficient to queue tasks on an idle node and migrate
to the new node. This would require additional smarts in the balancer so
for now the balancer will simply prefer to place the task on the preferred
node for a PTE scans which is controlled by the numa_balancing_settle_count
sysctl. Once the settle_count number of scans has complete the schedule
is free to place the task on an alternative node if the load is imbalanced.
[srikar@linux.vnet.ibm.com: Fixed statistics]
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
[ Tunable and use higher faults instead of preferred. ]
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-23-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
NUMA hinting fault counts and placement decisions are both recorded in the
same array which distorts the samples in an unpredictable fashion. The values
linearly accumulate during the scan and then decay creating a sawtooth-like
pattern in the per-node counts. It also means that placement decisions are
time sensitive. At best it means that it is very difficult to state that
the buffer holds a decaying average of past faulting behaviour. At worst,
it can confuse the load balancer if it sees one node with an artifically high
count due to very recent faulting activity and may create a bouncing effect.
This patch adds a second array. numa_faults stores the historical data
which is used for placement decisions. numa_faults_buffer holds the
fault activity during the current scan window. When the scan completes,
numa_faults decays and the values from numa_faults_buffer are copied
across.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-22-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This patch selects a preferred node for a task to run on based on the
NUMA hinting faults. This information is later used to migrate tasks
towards the node during balancing.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-21-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This patch tracks what nodes numa hinting faults were incurred on.
This information is later used to schedule a task on the node storing
the pages most frequently faulted by the task.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-20-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
NUMA PTE scanning slows if a NUMA hinting fault was trapped and no page
was migrated. For long-lived but idle processes there may be no faults
but the scan rate will be high and just waste CPU. This patch will slow
the scan rate for processes that are not trapping faults.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-19-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The NUMA PTE scan rate is controlled with a combination of the
numa_balancing_scan_period_min, numa_balancing_scan_period_max and
numa_balancing_scan_size. This scan rate is independent of the size
of the task and as an aside it is further complicated by the fact that
numa_balancing_scan_size controls how many pages are marked pte_numa and
not how much virtual memory is scanned.
In combination, it is almost impossible to meaningfully tune the min and
max scan periods and reasoning about performance is complex when the time
to complete a full scan is is partially a function of the tasks memory
size. This patch alters the semantic of the min and max tunables to be
about tuning the length time it takes to complete a scan of a tasks occupied
virtual address space. Conceptually this is a lot easier to understand. There
is a "sanity" check to ensure the scan rate is never extremely fast based on
the amount of virtual memory that should be scanned in a second. The default
of 2.5G seems arbitrary but it is to have the maximum scan rate after the
patch roughly match the maximum scan rate before the patch was applied.
On a similar note, numa_scan_period is in milliseconds and not
jiffies. Properly placed pages slow the scanning rate but adding 10 jiffies
to numa_scan_period means that the rate scanning slows depends on HZ which
is confusing. Get rid of the jiffies_to_msec conversion and treat it as ms.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-18-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Scan delay logic and resets are currently initialised to start scanning
immediately instead of delaying properly. Initialise them properly at
fork time and catch when a new mm has been allocated.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-17-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
PTE scanning and NUMA hinting fault handling is expensive so commit
5bca2303 ("mm: sched: numa: Delay PTE scanning until a task is scheduled
on a new node") deferred the PTE scan until a task had been scheduled on
another node. The problem is that in the purely shared memory case that
this may never happen and no NUMA hinting fault information will be
captured. We are not ruling out the possibility that something better
can be done here but for now, this patch needs to be reverted and depend
entirely on the scan_delay to avoid punishing short-lived processes.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-16-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Avoiding marking PTEs pte_numa because a particular NUMA node is migrate rate
limited sees like a bad idea. Even if this node can't migrate anymore other
nodes might and we want up-to-date information to do balance decisions.
We already rate limit the actual migrations, this should leave enough
bandwidth to allow the non-migrating scanning. I think its important we
keep up-to-date information if we're going to do placement based on it.
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Link: http://lkml.kernel.org/r/1381141781-10992-15-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
With a trace_printk("working\n"); right after the cmpxchg in
task_numa_work() we can see that of a 4 thread process, its always the
same task winning the race and doing the protection change.
This is a problem since the task doing the protection change has a
penalty for taking faults -- it is busy when marking the PTEs. If its
always the same task the ->numa_faults[] get severely skewed.
Avoid this by delaying the task doing the protection change such that
it is unlikely to win the privilege again.
Before:
root@interlagos:~# grep "thread 0/.*working" /debug/tracing/trace | tail -15
thread 0/0-3232 [022] .... 212.787402: task_numa_work: working
thread 0/0-3232 [022] .... 212.888473: task_numa_work: working
thread 0/0-3232 [022] .... 212.989538: task_numa_work: working
thread 0/0-3232 [022] .... 213.090602: task_numa_work: working
thread 0/0-3232 [022] .... 213.191667: task_numa_work: working
thread 0/0-3232 [022] .... 213.292734: task_numa_work: working
thread 0/0-3232 [022] .... 213.393804: task_numa_work: working
thread 0/0-3232 [022] .... 213.494869: task_numa_work: working
thread 0/0-3232 [022] .... 213.596937: task_numa_work: working
thread 0/0-3232 [022] .... 213.699000: task_numa_work: working
thread 0/0-3232 [022] .... 213.801067: task_numa_work: working
thread 0/0-3232 [022] .... 213.903155: task_numa_work: working
thread 0/0-3232 [022] .... 214.005201: task_numa_work: working
thread 0/0-3232 [022] .... 214.107266: task_numa_work: working
thread 0/0-3232 [022] .... 214.209342: task_numa_work: working
After:
root@interlagos:~# grep "thread 0/.*working" /debug/tracing/trace | tail -15
thread 0/0-3253 [005] .... 136.865051: task_numa_work: working
thread 0/2-3255 [026] .... 136.965134: task_numa_work: working
thread 0/3-3256 [024] .... 137.065217: task_numa_work: working
thread 0/3-3256 [024] .... 137.165302: task_numa_work: working
thread 0/3-3256 [024] .... 137.265382: task_numa_work: working
thread 0/0-3253 [004] .... 137.366465: task_numa_work: working
thread 0/2-3255 [026] .... 137.466549: task_numa_work: working
thread 0/0-3253 [004] .... 137.566629: task_numa_work: working
thread 0/0-3253 [004] .... 137.666711: task_numa_work: working
thread 0/1-3254 [028] .... 137.766799: task_numa_work: working
thread 0/0-3253 [004] .... 137.866876: task_numa_work: working
thread 0/2-3255 [026] .... 137.966960: task_numa_work: working
thread 0/1-3254 [028] .... 138.067041: task_numa_work: working
thread 0/2-3255 [026] .... 138.167123: task_numa_work: working
thread 0/3-3256 [024] .... 138.267207: task_numa_work: working
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Link: http://lkml.kernel.org/r/1381141781-10992-14-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Fix a 80 column violation and a PTE vs PMD reference.
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Link: http://lkml.kernel.org/r/1381141781-10992-4-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
In 76854c7e8f ("sched: Use
rt.nr_cpus_allowed to recover select_task_rq() cycles") an
optimization was added to select_task_rq_rt() that immediately
returns when p->nr_cpus_allowed == 1 at the beginning of the
function.
This makes the latter p->nr_cpus_allowed > 1 check redundant,
which can now be removed.
Signed-off-by: Shawn Bohrer <sbohrer@rgmadvisors.com>
Reviewed-by: Steven Rostedt <rostedt@goodmis.org>
Cc: Mike Galbraith <mgalbraith@suse.de>
Cc: tomk@rgmadvisors.com
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1380914693-24634-1-git-send-email-shawn.bohrer@gmail.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
When using per-cpu preempt_count variables we need to save/restore the
preempt_count on context switch (into per task storage; for instance
the old thread_info::preempt_count variable) because of
PREEMPT_ACTIVE.
However, this means that on fork() the preempt_count value of the last
context switch gets copied and if we had a PREEMPT_ACTIVE switch right
before cloning a child task the child task will now too have
PREEMPT_ACTIVE set and start its life with an extra PREEMPT_ACTIVE
count.
Therefore we need to make init_task_preempt_count() unconditional;
this resets whatever preempt_count we inherited from our parent
process.
Doing so for !per-cpu implementations is harmless.
For !PREEMPT_COUNT kernels we need to be careful not to start life
with an increased preempt_count.
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-4k0b7oy1rcdyzochwiixuwi9@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Rewrite the preempt_count macros in order to extract the 3 basic
preempt_count value modifiers:
__preempt_count_add()
__preempt_count_sub()
and the new:
__preempt_count_dec_and_test()
And since we're at it anyway, replace the unconventional
$op_preempt_count names with the more conventional preempt_count_$op.
Since these basic operators are equivalent to the previous _notrace()
variants, do away with the _notrace() versions.
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-ewbpdbupy9xpsjhg960zwbv8@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
We need a few special preempt_count accessors:
- task_preempt_count() for when we're interested in the preemption
count of another (non-running) task.
- init_task_preempt_count() for properly initializing the preemption
count.
- init_idle_preempt_count() a special case of the above for the idle
threads.
With these no generic code ever touches thread_info::preempt_count
anymore and architectures could choose to remove it.
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-jf5swrio8l78j37d06fzmo4r@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
In order to combine the preemption and need_resched test we need to
fold the need_resched information into the preempt_count value.
Since the NEED_RESCHED flag is set across CPUs this needs to be an
atomic operation, however we very much want to avoid making
preempt_count atomic, therefore we keep the existing TIF_NEED_RESCHED
infrastructure in place but at 3 sites test it and fold its value into
preempt_count; namely:
- resched_task() when setting TIF_NEED_RESCHED on the current task
- scheduler_ipi() when resched_task() sets TIF_NEED_RESCHED on a
remote task it follows it up with a reschedule IPI
and we can modify the cpu local preempt_count from
there.
- cpu_idle_loop() for when resched_task() found tsk_is_polling().
We use an inverted bitmask to indicate need_resched so that a 0 means
both need_resched and !atomic.
Also remove the barrier() in preempt_enable() between
preempt_enable_no_resched() and preempt_check_resched() to avoid
having to reload the preemption value and allow the compiler to use
the flags of the previuos decrement. I couldn't come up with any sane
reason for this barrier() to be there as preempt_enable_no_resched()
already has a barrier() before doing the decrement.
Suggested-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-7a7m5qqbn5pmwnd4wko9u6da@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Replace the single preempt_count() 'function' that's an lvalue with
two proper functions:
preempt_count() - returns the preempt_count value as rvalue
preempt_count_set() - Allows setting the preempt-count value
Also provide preempt_count_ptr() as a convenience wrapper to implement
all modifying operations.
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-orxrbycjozopqfhb4dxdkdvb@git.kernel.org
[ Fixed build failure. ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
We're going to deprecate and remove set_need_resched() for it will do
the wrong thing. Make an exception for RCU and allow it to use
resched_cpu() which will do the right thing.
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Cc: Paul McKenney <paulmck@linux.vnet.ibm.com>
Link: http://lkml.kernel.org/n/tip-2eywnacjl1nllctl1nszqa5w@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
We always know the rq used, let's just pass it around.
This seems to cut the size of scheduler core down a tiny bit:
Before:
[linux]$ size kernel/sched/core.o.orig
text data bss dec hex filename
62760 16130 3876 82766 1434e kernel/sched/core.o.orig
After:
[linux]$ size kernel/sched/core.o.patched
text data bss dec hex filename
62566 16130 3876 82572 1428c kernel/sched/core.o.patched
Probably speeds it up as well.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20130922142054.GA11499@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This patch builds on patch 2 and periodically decays that max value to
do idle balancing per sched domain by approximately 1% per second. Also
decay the rq's max_idle_balance_cost value.
Signed-off-by: Jason Low <jason.low2@hp.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1379096813-3032-4-git-send-email-jason.low2@hp.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
In this patch, we keep track of the max cost we spend doing idle load balancing
for each sched domain. If the avg time the CPU remains idle is less then the
time we have already spent on idle balancing + the max cost of idle balancing
in the sched domain, then we don't continue to attempt the balance. We also
keep a per rq variable, max_idle_balance_cost, which keeps track of the max
time spent on newidle load balances throughout all its domains so that we can
determine the avg_idle's max value.
By using the max, we avoid overrunning the average. This further reduces the
chance we attempt balancing when the CPU is not idle for longer than the cost
to balance.
Signed-off-by: Jason Low <jason.low2@hp.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1379096813-3032-3-git-send-email-jason.low2@hp.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
When updating avg_idle, if the delta exceeds some max value, then avg_idle
gets set to the max, regardless of what the previous avg was. This can cause
avg_idle to often be overestimated.
This patch modifies the way we update avg_idle by always updating it with the
function call to update_avg() first. Then, if avg_idle exceeds the max, we set
it to the max.
Signed-off-by: Jason Low <jason.low2@hp.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Reviewed-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1379096813-3032-2-git-send-email-jason.low2@hp.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Patch a003a2 (sched: Consider runnable load average in move_tasks())
sets all top-level cfs_rqs' h_load to rq->avg.load_avg_contrib, which is
always 0. This mistype leads to all tasks having weight 0 when load
balancing in a cpu-cgroup enabled setup. There obviously should be sum
of weights of all runnable tasks there instead. Fix it.
Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Reviewed-by: Paul Turner <pjt@google.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1379173186-11944-1-git-send-email-vdavydov@parallels.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
In busiest->group_imb case we can come to fix_small_imbalance() with
local->avg_load > busiest->avg_load. This can result in wrong imbalance
fix-up, because there is the following check there where all the
members are unsigned:
if (busiest->avg_load - local->avg_load + scaled_busy_load_per_task >=
(scaled_busy_load_per_task * imbn)) {
env->imbalance = busiest->load_per_task;
return;
}
As a result we can end up constantly bouncing tasks from one cpu to
another if there are pinned tasks.
Fix it by substituting the subtraction with an equivalent addition in
the check.
[ The bug can be caught by running 2*N cpuhogs pinned to two logical cpus
belonging to different cores on an HT-enabled machine with N logical
cpus: just look at se.nr_migrations growth. ]
Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/ef167822e5c5b2d96cf5b0e3e4f4bdff3f0414a2.1379252740.git.vdavydov@parallels.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
In busiest->group_imb case we can come to calculate_imbalance() with
local->avg_load >= busiest->avg_load >= sds->avg_load. This can result
in imbalance overflow, because it is calculated as follows
env->imbalance = min(
max_pull * busiest->group_power,
(sds->avg_load - local->avg_load) * local->group_power) / SCHED_POWER_SCALE;
As a result we can end up constantly bouncing tasks from one cpu to
another if there are pinned tasks.
Fix this by skipping the assignment and assuming imbalance=0 in case
local->avg_load > sds->avg_load.
[ The bug can be caught by running 2*N cpuhogs pinned to two logical cpus
belonging to different cores on an HT-enabled machine with N logical
cpus: just look at se.nr_migrations growth. ]
Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/8f596cc6bc0e5e655119dc892c9bfcad26e971f4.1379252740.git.vdavydov@parallels.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull scheduler fixes from Ingo Molnar:
"Misc fixes"
* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched: Fix comment for sched_info_depart
sched/Documentation: Update sched-design-CFS.txt documentation
sched/debug: Take PID namespace into account
sched/fair: Fix small race where child->se.parent,cfs_rq might point to invalid ones
sched_info_depart seems to be only called from
sched_info_switch(), so only on involuntary task switch.
Fix the comment to match.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Link: http://lkml.kernel.org/r/20130916083036.GA1113@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Do away with 'phantom' cores due to N*frac(smt_power) >= 1 by limiting
the capacity to the actual number of cores.
The assumption of 1 < smt_power < 2 is an actual requirement because
of what SMT is so this should work regardless of the SMT
implementation.
It can still be defeated by creative use of cpu hotplug, but if you're
one of those freaks, you get to live with it.
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Acked-by: Vincent Guittot <vincent.guitto@linaro.org>
Link: http://lkml.kernel.org/n/tip-dczmbi8tfgixacg1ji2av1un@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
When looking at the code I noticed we don't actually compute
sgp->power_orig correctly for groups, fix that.
Currently the only consumer of that value is fix_small_capacity()
which is only used on POWER7+ and that code excludes this case by
being limited to SD_SHARE_CPUPOWER which is only ever set on the SMT
domain which must be the lowest domain and this has singleton groups.
So nothing should be affected by this change.
Cc: Michael Neuling <mikey@neuling.org>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-db2pe0vxwunv37plc7onnugj@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Change the group_imb detection from the old 'load-spike' detector to
an actual imbalance detector. We set it from the lower domain balance
pass when it fails to create a balance in the presence of task
affinities.
The advantage is that this should no longer generate the false
positive group_imb conditions generated by transient load spikes from
the normal balancing/bulk-wakeup etc. behaviour.
While I haven't actually observed those they could happen.
I'm not entirely happy with this patch; it somehow feels a little
fragile.
Nor does it solve the biggest issue I have with the group_imb code; it
it still a fragile construct in that once we 'fixed' the imbalance
we'll not detect the group_imb again and could end up re-creating it.
That said, this patch does seem to preserve behaviour for the
described degenerate case. In particular on my 2*6*2 wsm-ep:
taskset -c 3-11 bash -c 'for ((i=0;i<9;i++)) do while :; do :; done & done'
ends up with 9 spinners, each on their own CPU; whereas if you disable
the group_imb code that typically doesn't happen (you'll get one pair
sharing a CPU most of the time).
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-36fpbgl39dv4u51b6yz2ypz5@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Emmanuel reported that /proc/sched_debug didn't report the right PIDs
when using namespaces, cure this.
Reported-by: Emmanuel Deloget <emmanuel.deloget@efixo.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20130909110141.GM31370@twins.programming.kicks-ass.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
There is a small race between copy_process() and cgroup_attach_task()
where child->se.parent,cfs_rq points to invalid (old) ones.
parent doing fork() | someone moving the parent to another cgroup
-------------------------------+---------------------------------------------
copy_process()
+ dup_task_struct()
-> parent->se is copied to child->se.
se.parent,cfs_rq of them point to old ones.
cgroup_attach_task()
+ cgroup_task_migrate()
-> parent->cgroup is updated.
+ cpu_cgroup_attach()
+ sched_move_task()
+ task_move_group_fair()
+- set_task_rq()
-> se.parent,cfs_rq of parent
are updated.
+ cgroup_fork()
-> parent->cgroup is copied to child->cgroup. (*1)
+ sched_fork()
+ task_fork_fair()
-> se.parent,cfs_rq of child are accessed
while they point to old ones. (*2)
In the worst case, this bug can lead to "use-after-free" and cause a panic,
because it's new cgroup's refcount that is incremented at (*1),
so the old cgroup(and related data) can be freed before (*2).
In fact, a panic caused by this bug was originally caught in RHEL6.4.
BUG: unable to handle kernel NULL pointer dereference at (null)
IP: [<ffffffff81051e3e>] sched_slice+0x6e/0xa0
[...]
Call Trace:
[<ffffffff81051f25>] place_entity+0x75/0xa0
[<ffffffff81056a3a>] task_fork_fair+0xaa/0x160
[<ffffffff81063c0b>] sched_fork+0x6b/0x140
[<ffffffff8106c3c2>] copy_process+0x5b2/0x1450
[<ffffffff81063b49>] ? wake_up_new_task+0xd9/0x130
[<ffffffff8106d2f4>] do_fork+0x94/0x460
[<ffffffff81072a9e>] ? sys_wait4+0xae/0x100
[<ffffffff81009598>] sys_clone+0x28/0x30
[<ffffffff8100b393>] stub_clone+0x13/0x20
[<ffffffff8100b072>] ? system_call_fastpath+0x16/0x1b
Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Cc: <stable@vger.kernel.org>
Link: http://lkml.kernel.org/r/039601ceae06$733d3130$59b79390$@mxp.nes.nec.co.jp
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Commit 23f0d20 ("sched: Factor out code to should_we_balance()")
introduces the should_we_balance() function. This function should
return 1 if this cpu is appropriate for balancing. But the newly
introduced code doesn't do so, it returns 0 instead of 1.
This introduces performance regression, reported by Dave Chinner:
v4 filesystem v5 filesystem
3.11+xfsdev: 220k files/s 225k files/s
3.12-git 180k files/s 185k files/s
3.12-git-revert 245k files/s 247k files/s
You can find more detailed information at:
https://lkml.org/lkml/2013/9/10/1
This patch corrects the return value of should_we_balance()
function as orignally intended.
With this patch, Dave Chinner reports that the regression is gone:
v4 filesystem v5 filesystem
3.11+xfsdev: 220k files/s 225k files/s
3.12-git 180k files/s 185k files/s
3.12-git-revert 245k files/s 247k files/s
3.12-git-fix 249k files/s 248k files/s
Reported-by: Dave Chinner <dchinner@redhat.com>
Tested-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Paul Turner <pjt@google.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Dave Chinner <david@fromorbit.com>
Link: http://lkml.kernel.org/r/20130910065448.GA20368@lge.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull cputime fix from Ingo Molnar:
"This fixes a longer-standing cputime accounting bug that Stanislaw
Gruszka finally managed to track down"
* 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched/cputime: Do not scale when utime == 0
Pull KVM updates from Gleb Natapov:
"The highlights of the release are nested EPT and pv-ticketlocks
support (hypervisor part, guest part, which is most of the code, goes
through tip tree). Apart of that there are many fixes for all arches"
Fix up semantic conflicts as discussed in the pull request thread..
* 'next' of git://git.kernel.org/pub/scm/virt/kvm/kvm: (88 commits)
ARM: KVM: Add newlines to panic strings
ARM: KVM: Work around older compiler bug
ARM: KVM: Simplify tracepoint text
ARM: KVM: Fix kvm_set_pte assignment
ARM: KVM: vgic: Bump VGIC_NR_IRQS to 256
ARM: KVM: Bugfix: vgic_bytemap_get_reg per cpu regs
ARM: KVM: vgic: fix GICD_ICFGRn access
ARM: KVM: vgic: simplify vgic_get_target_reg
KVM: MMU: remove unused parameter
KVM: PPC: Book3S PR: Rework kvmppc_mmu_book3s_64_xlate()
KVM: PPC: Book3S PR: Make instruction fetch fallback work for system calls
KVM: PPC: Book3S PR: Don't corrupt guest state when kernel uses VMX
KVM: x86: update masterclock when kvmclock_offset is calculated (v2)
KVM: PPC: Book3S: Fix compile error in XICS emulation
KVM: PPC: Book3S PR: return appropriate error when allocation fails
arch: powerpc: kvm: add signed type cast for comparation
KVM: x86: add comments where MMIO does not return to the emulator
KVM: vmx: count exits to userspace during invalid guest emulation
KVM: rename __kvm_io_bus_sort_cmp to kvm_io_bus_cmp
kvm: optimize away THP checks in kvm_is_mmio_pfn()
...
Pull timers/nohz changes from Ingo Molnar:
"It mostly contains fixes and full dynticks off-case optimizations, by
Frederic Weisbecker"
* 'timers-nohz-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (24 commits)
nohz: Include local CPU in full dynticks global kick
nohz: Optimize full dynticks's sched hooks with static keys
nohz: Optimize full dynticks state checks with static keys
nohz: Rename a few state variables
vtime: Always debug check snapshot source _before_ updating it
vtime: Always scale generic vtime accounting results
vtime: Optimize full dynticks accounting off case with static keys
vtime: Describe overriden functions in dedicated arch headers
m68k: hardirq_count() only need preempt_mask.h
hardirq: Split preempt count mask definitions
context_tracking: Split low level state headers
vtime: Fix racy cputime delta update
vtime: Remove a few unneeded generic vtime state checks
context_tracking: User/kernel broundary cross trace events
context_tracking: Optimize context switch off case with static keys
context_tracking: Optimize guest APIs off case with static key
context_tracking: Optimize main APIs off case with static key
context_tracking: Ground setup for static key use
context_tracking: Remove full dynticks' hacky dependency on wide context tracking
nohz: Only enable context tracking on full dynticks CPUs
...
Pull scheduler changes from Ingo Molnar:
"Various optimizations, cleanups and smaller fixes - no major changes
in scheduler behavior"
* 'sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched/fair: Fix the sd_parent_degenerate() code
sched/fair: Rework and comment the group_imb code
sched/fair: Optimize find_busiest_queue()
sched/fair: Make group power more consistent
sched/fair: Remove duplicate load_per_task computations
sched/fair: Shrink sg_lb_stats and play memset games
sched: Clean-up struct sd_lb_stat
sched: Factor out code to should_we_balance()
sched: Remove one division operation in find_busiest_queue()
sched/cputime: Use this_cpu_add() in task_group_account_field()
cpumask: Fix cpumask leak in partition_sched_domains()
sched/x86: Optimize switch_mm() for multi-threaded workloads
generic-ipi: Kill unnecessary variable - csd_flags
numa: Mark __node_set() as __always_inline
sched/fair: Cleanup: remove duplicate variable declaration
sched/__wake_up_sync_key(): Fix nr_exclusive tasks which lead to WF_SYNC clearing
Pull perf changes from Ingo Molnar:
"As a first remark I'd like to point out that the obsolete '-f'
(--force) option, which has not done anything for several releases,
has been removed from 'perf record' and related utilities. Everyone
please update muscle memory accordingly! :-)
Main changes on the perf kernel side:
- Performance optimizations:
. for trace events, by Steve Rostedt.
. for time values, by Peter Zijlstra
- New hardware support:
. for Intel Silvermont (22nm Atom) CPUs, by Zheng Yan
. for Intel SNB-EP uncore PMUs, by Zheng Yan
- Enhanced hardware support:
. for Intel uncore PMUs: add filter support for QPI boxes, by Zheng Yan
- Core perf events code enhancements and fixes:
. for full-nohz feature handling, by Frederic Weisbecker
. for group events, by Jiri Olsa
. for call chains, by Frederic Weisbecker
. for event stream parsing, by Adrian Hunter
- New ABI details:
. Add attr->mmap2 attribute, by Stephane Eranian
. Add PERF_EVENT_IOC_ID ioctl to return event ID, by Jiri Olsa
. Export u64 time_zero on the mmap header page to allow TSC
calculation, by Adrian Hunter
. Add dummy software event, by Adrian Hunter.
. Add a new PERF_SAMPLE_IDENTIFIER to make samples always
parseable, by Adrian Hunter.
. Make Power7 events available via sysfs, by Runzhen Wang.
- Code cleanups and refactorings:
. for nohz-full, by Frederic Weisbecker
. for group events, by Jiri Olsa
- Documentation updates:
. for perf_event_type, by Peter Zijlstra
Main changes on the perf tooling side (some of these tooling changes
utilize the above kernel side changes):
- Lots of 'perf trace' enhancements:
. Make 'perf trace' command line arguments consistent with
'perf record', by David Ahern.
. Allow specifying syscalls a la strace, by Arnaldo Carvalho de Melo.
. Add --verbose and -o/--output options, by Arnaldo Carvalho de Melo.
. Support ! in -e expressions, to filter a list of syscalls,
by Arnaldo Carvalho de Melo.
. Arg formatting improvements to allow masking arguments in
syscalls such as futex and open, where the some arguments are
ignored and thus should not be printed depending on other args,
by Arnaldo Carvalho de Melo.
. Beautify futex open, openat, open_by_handle_at, lseek and futex
syscalls, by Arnaldo Carvalho de Melo.
. Add option to analyze events in a file versus live, so that
one can do:
[root@zoo ~]# perf record -a -e raw_syscalls:* sleep 1
[ perf record: Woken up 0 times to write data ]
[ perf record: Captured and wrote 25.150 MB perf.data (~1098836 samples) ]
[root@zoo ~]# perf trace -i perf.data -e futex --duration 1
17.799 ( 1.020 ms): 7127 futex(uaddr: 0x7fff3f6c6674, op: 393, val: 1, utime: 0x7fff3f6c6470, ua
113.344 (95.429 ms): 7127 futex(uaddr: 0x7fff3f6c6674, op: 393, val: 1, utime: 0x7fff3f6c6470, uaddr2: 0x7fff3f6c6648, val3: 4294967
133.778 ( 1.042 ms): 18004 futex(uaddr: 0x7fff3f6c6674, op: 393, val: 1, utime: 0x7fff3f6c6470, uaddr2: 0x7fff3f6c6648, val3: 429496
[root@zoo ~]#
By David Ahern.
. Honor target pid / tid options when analyzing a file, by David Ahern.
. Introduce better formatting of syscall arguments, including so
far beautifiers for mmap, madvise, syscall return values,
by Arnaldo Carvalho de Melo.
. Handle HUGEPAGE defines in the mmap beautifier, by David Ahern.
- 'perf report/top' enhancements:
. Do annotation using /proc/kcore and /proc/kallsyms when
available, removing the forced need for a vmlinux file kernel
assembly annotation. This also improves this use case because
vmlinux has just the initial kernel image, not what is actually
in use after various code patchings by things like alternatives.
By Adrian Hunter.
. Add --ignore-callees=<regex> option to collapse undesired parts
of call graphs, by Greg Price.
. Simplify symbol filtering by doing it at machine class level,
by Adrian Hunter.
. Add support for callchains in the gtk UI, by Namhyung Kim.
. Add --objdump option to 'perf top', by Sukadev Bhattiprolu.
- 'perf kvm' enhancements:
. Add option to print only events that exceed a specified time
duration, by David Ahern.
. Improve stack trace printing, by David Ahern.
. Update documentation of the live command, by David Ahern
. Add perf kvm stat live mode that combines aspects of 'perf kvm
stat' record and report, by David Ahern.
. Add option to analyze specific VM in perf kvm stat report, by
David Ahern.
. Do not require /lib/modules/* on a guest, by Jason Wessel.
- 'perf script' enhancements:
. Fix symbol offset computation for some dsos, by David Ahern.
. Fix named threads support, by David Ahern.
. Don't install scripting files files when perl/python support
is disabled, by Arnaldo Carvalho de Melo.
- 'perf test' enhancements:
. Add various improvements and fixes to the "vmlinux matches
kallsyms" 'perf test' entry, related to the /proc/kcore
annotation feature. By Adrian Hunter.
. Add sample parsing test, by Adrian Hunter.
. Add test for reading object code, by Adrian Hunter.
. Add attr record group sampling test, by Jiri Olsa.
. Misc testing infrastructure improvements and other details,
by Jiri Olsa.
- 'perf list' enhancements:
. Skip unsupported hardware events, by Namhyung Kim.
. List pmu events, by Andi Kleen.
- 'perf diff' enhancements:
. Add support for more than two files comparison, by Jiri Olsa.
- 'perf sched' enhancements:
. Various improvements, including removing reliance on some
scheduler tracepoints that provide the same information as the
PERF_RECORD_{FORK,EXIT} events. By David Ahern.
. Remove odd build stall by moving a large struct initialization
from a local variable to a global one, by Namhyung Kim.
- 'perf stat' enhancements:
. Add --initial-delay option to skip measuring for a defined
startup phase, by Andi Kleen.
- Generic perf tooling infrastructure/plumbing changes:
. Tidy up sample parsing validation, by Adrian Hunter.
. Fix up jobserver setup in libtraceevent Makefile.
by Arnaldo Carvalho de Melo.
. Debug improvements, by Adrian Hunter.
. Fix correlation of samples coming after PERF_RECORD_EXIT event,
by David Ahern.
. Improve robustness of the topology parsing code,
by Stephane Eranian.
. Add group leader sampling, that allows just one event in a group
to sample while the other events have just its values read,
by Jiri Olsa.
. Add support for a new modifier "D", which requests that the
event, or group of events, be pinned to the PMU.
By Michael Ellerman.
. Support callchain sorting based on addresses, by Andi Kleen
. Prep work for multi perf data file storage, by Jiri Olsa.
. libtraceevent cleanups, by Namhyung Kim.
And lots and lots of other fixes and code reorganizations that did not
make it into the list, see the shortlog, diffstat and the Git log for
details!"
[ Also merge a leftover from the 3.11 cycle ]
* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf: Prevent race in unthrottling code
* 'perf-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (237 commits)
perf trace: Tell arg formatters the arg index
perf trace: Add beautifier for open's flags arg
perf trace: Add beautifier for lseek's whence arg
perf tools: Fix symbol offset computation for some dsos
perf list: Skip unsupported events
perf tests: Add 'keep tracking' test
perf tools: Add support for PERF_COUNT_SW_DUMMY
perf: Add a dummy software event to keep tracking
perf trace: Add beautifier for futex 'operation' parm
perf trace: Allow syscall arg formatters to mask args
perf: Convert kmalloc_node(...GFP_ZERO...) to kzalloc_node()
perf: Export struct perf_branch_entry to userspace
perf: Add attr->mmap2 attribute to an event
perf/x86: Add Silvermont (22nm Atom) support
perf/x86: use INTEL_UEVENT_EXTRA_REG to define MSR_OFFCORE_RSP_X
perf trace: Handle missing HUGEPAGE defines
perf trace: Honor target pid / tid options when analyzing a file
perf trace: Add option to analyze events in a file versus live
perf evlist: Add tracepoint lookup by name
perf tests: Add a sample parsing test
...
scale_stime() silently assumes that stime < rtime, otherwise
when stime == rtime and both values are big enough (operations
on them do not fit in 32 bits), the resulting scaling stime can
be bigger than rtime. In consequence utime = rtime - stime
results in negative value.
User space visible symptoms of the bug are overflowed TIME
values on ps/top, for example:
$ ps aux | grep rcu
root 8 0.0 0.0 0 0 ? S 12:42 0:00 [rcuc/0]
root 9 0.0 0.0 0 0 ? S 12:42 0:00 [rcub/0]
root 10 62422329 0.0 0 0 ? R 12:42 21114581:37 [rcu_preempt]
root 11 0.1 0.0 0 0 ? S 12:42 0:02 [rcuop/0]
root 12 62422329 0.0 0 0 ? S 12:42 21114581:35 [rcuop/1]
root 10 62422329 0.0 0 0 ? R 12:42 21114581:37 [rcu_preempt]
or overflowed utime values read directly from /proc/$PID/stat
Reference:
https://lkml.org/lkml/2013/8/20/259
Reported-and-tested-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Cc: stable@vger.kernel.org
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Borislav Petkov <bp@alien8.de>
Link: http://lkml.kernel.org/r/20130904131602.GC2564@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull cgroup updates from Tejun Heo:
"A lot of activities on the cgroup front. Most changes aren't visible
to userland at all at this point and are laying foundation for the
planned unified hierarchy.
- The biggest change is decoupling the lifetime management of css
(cgroup_subsys_state) from that of cgroup's. Because controllers
(cpu, memory, block and so on) will need to be dynamically enabled
and disabled, css which is the association point between a cgroup
and a controller may come and go dynamically across the lifetime of
a cgroup. Till now, css's were created when the associated cgroup
was created and stayed till the cgroup got destroyed.
Assumptions around this tight coupling permeated through cgroup
core and controllers. These assumptions are gradually removed,
which consists bulk of patches, and css destruction path is
completely decoupled from cgroup destruction path. Note that
decoupling of creation path is relatively easy on top of these
changes and the patchset is pending for the next window.
- cgroup has its own event mechanism cgroup.event_control, which is
only used by memcg. It is overly complex trying to achieve high
flexibility whose benefits seem dubious at best. Going forward,
new events will simply generate file modified event and the
existing mechanism is being made specific to memcg. This pull
request contains prepatory patches for such change.
- Various fixes and cleanups"
Fixed up conflict in kernel/cgroup.c as per Tejun.
* 'for-3.12' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup: (69 commits)
cgroup: fix cgroup_css() invocation in css_from_id()
cgroup: make cgroup_write_event_control() use css_from_dir() instead of __d_cgrp()
cgroup: make cgroup_event hold onto cgroup_subsys_state instead of cgroup
cgroup: implement CFTYPE_NO_PREFIX
cgroup: make cgroup_css() take cgroup_subsys * instead and allow NULL subsys
cgroup: rename cgroup_css_from_dir() to css_from_dir() and update its syntax
cgroup: fix cgroup_write_event_control()
cgroup: fix subsystem file accesses on the root cgroup
cgroup: change cgroup_from_id() to css_from_id()
cgroup: use css_get() in cgroup_create() to check CSS_ROOT
cpuset: remove an unncessary forward declaration
cgroup: RCU protect each cgroup_subsys_state release
cgroup: move subsys file removal to kill_css()
cgroup: factor out kill_css()
cgroup: decouple cgroup_subsys_state destruction from cgroup destruction
cgroup: replace cgroup->css_kill_cnt with ->nr_css
cgroup: bounce cgroup_subsys_state ref kill confirmation to a work item
cgroup: move cgroup->subsys[] assignment to online_css()
cgroup: reorganize css init / exit paths
cgroup: add __rcu modifier to cgroup->subsys[]
...
I found that on my WSM box I had a redundant domain:
[ 0.949769] CPU0 attaching sched-domain:
[ 0.953765] domain 0: span 0,12 level SIBLING
[ 0.958335] groups: 0 (cpu_power = 587) 12 (cpu_power = 588)
[ 0.964548] domain 1: span 0-5,12-17 level MC
[ 0.969206] groups: 0,12 (cpu_power = 1175) 1,13 (cpu_power = 1176) 2,14 (cpu_power = 1176) 3,15 (cpu_power = 1176) 4,16 (cpu_power = 1176) 5,17 (cpu_power = 1176)
[ 0.984993] domain 2: span 0-5,12-17 level CPU
[ 0.989822] groups: 0-5,12-17 (cpu_power = 7055)
[ 0.995049] domain 3: span 0-23 level NUMA
[ 0.999620] groups: 0-5,12-17 (cpu_power = 7055) 6-11,18-23 (cpu_power = 7056)
Note how domain 2 has only a single group and spans the same CPUs as
domain 1. We should not keep such domains and do in fact have code to
prune these.
It turns out that the 'new' SD_PREFER_SIBLING flag causes this, it
makes sd_parent_degenerate() fail on the CPU domain. We can easily
fix this by 'ignoring' the SD_PREFER_SIBLING bit and transfering it
to whatever domain ends up covering the span.
With this patch the domains now look like this:
[ 0.950419] CPU0 attaching sched-domain:
[ 0.954454] domain 0: span 0,12 level SIBLING
[ 0.959039] groups: 0 (cpu_power = 587) 12 (cpu_power = 588)
[ 0.965271] domain 1: span 0-5,12-17 level MC
[ 0.969936] groups: 0,12 (cpu_power = 1175) 1,13 (cpu_power = 1176) 2,14 (cpu_power = 1176) 3,15 (cpu_power = 1176) 4,16 (cpu_power = 1176) 5,17 (cpu_power = 1176)
[ 0.985737] domain 2: span 0-23 level NUMA
[ 0.990231] groups: 0-5,12-17 (cpu_power = 7055) 6-11,18-23 (cpu_power = 7056)
Reviewed-by: Paul Turner <pjt@google.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-ys201g4jwukj0h8xcamakxq1@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Rik reported some weirdness due to the group_imb code. As a start to
looking at it, clean it up a little and add a few explanatory
comments.
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-caeeqttnla4wrrmhp5uf89gp@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Use for_each_cpu_and() and thereby avoid computing the capacity for
CPUs we know we're not interested in.
Reviewed-by: Paul Turner <pjt@google.com>
Reviewed-by: Preeti U Murthy <preeti@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-lppceyv6kb3a19g8spmrn20b@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
For easier access, less dereferences and more consistent value, store
the group power in update_sg_lb_stats() and use it thereafter. The
actual value in sched_group::sched_group_power::power can change
throughout the load-balance pass if we're unlucky.
Reviewed-by: Preeti U Murthy <preeti@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-739xxqkyvftrhnh9ncudutc7@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Since we already compute (but don't store) the sgs load_per_task value
in update_sg_lb_stats() we might as well store it and not re-compute
it later on.
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-ym1vmljiwbzgdnnrwp9azftq@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
We can shrink sg_lb_stats because rq::nr_running is an unsigned int
and cpu numbers are 'int'
Before:
sgs: /* size: 72, cachelines: 2, members: 10 */
sds: /* size: 184, cachelines: 3, members: 7 */
After:
sgs: /* size: 56, cachelines: 1, members: 10 */
sds: /* size: 152, cachelines: 3, members: 7 */
Further we can avoid clearing all of sds since we do a total
clear/assignment of sg_stats in update_sg_lb_stats() with exception of
busiest_stat.avg_load which is referenced in update_sd_pick_busiest().
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-0klzmz9okll8wc0nsudguc9p@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
There is no reason to maintain separate variables for this_group
and busiest_group in sd_lb_stat, except saving some space.
But this structure is always allocated in stack, so this saving
isn't really benificial [peterz: reducing stack space is good; in this
case readability increases enough that I think its still beneficial]
This patch unify these variables, so IMO, readability may be improved.
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
[ Rename this to local -- avoids confusion between this_cpu and the C++ this pointer. ]
Reviewed-by: Paul Turner <pjt@google.com>
[ Lots of style edits, a few fixes and a rename. ]
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1375778203-31343-4-git-send-email-iamjoonsoo.kim@lge.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Now checking whether this cpu is appropriate to balance or not
is embedded into update_sg_lb_stats() and this checking has no direct
relationship to this function. There is not enough reason to place
this checking at update_sg_lb_stats(), except saving one iteration
for sched_group_cpus.
In this patch, I factor out this checking to should_we_balance() function.
And before doing actual work for load_balancing, check whether this cpu is
appropriate to balance via should_we_balance(). If this cpu is not
a candidate for balancing, it quit the work immediately.
With this change, we can save two memset cost and can expect better
compiler optimization.
Below is result of this patch.
* Vanilla *
text data bss dec hex filename
34499 1136 116 35751 8ba7 kernel/sched/fair.o
* Patched *
text data bss dec hex filename
34243 1136 116 35495 8aa7 kernel/sched/fair.o
In addition, rename @balance to @continue_balancing in order to represent
its purpose more clearly.
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
[ s/should_balance/continue_balancing/g ]
Reviewed-by: Paul Turner <pjt@google.com>
[ Made style changes and a fix in should_we_balance(). ]
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1375778203-31343-3-git-send-email-iamjoonsoo.kim@lge.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Use of a this_cpu() operation reduces the number of instructions used
for accounting (account_user_time()) and frees up some registers. This is in
the scheduler tick hotpath.
Signed-off-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/00000140596dd165-338ff7f5-893b-4fec-b251-aaac5557239e-000000@email.amazonses.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
If doms_new is NULL, partition_sched_domains() will reset ndoms_cur
to 0, and free old sched domains with free_sched_domains(doms_cur, ndoms_cur).
As ndoms_cur is 0, the cpumask will not be freed.
Signed-off-by: Xiaotian Feng <xtfeng@gmail.com>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1375790802-11857-1-git-send-email-xtfeng@gmail.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)
iQEcBAABAgAGBQJSCDSjAAoJEHm+PkMAQRiGDXMIAI7Loae0Oqb1eoeJkvjyZsBS
OJDeeEcn+k58VbxVHyRdc7hGo4yI4tUZm172SpnOaM8sZ/ehPU7zBrwJK2lzX334
/jAM3uvVPfxA2nu0I4paNpkED/NQ8NRRsYE1iTE8dzHXOH6dA3mgp5qfco50rQvx
rvseXpME4KIAJEq4jnyFZF5+nuHiPueM9JftPmSSmJJ3/KY9kY1LESovyWd7ttg1
jYSVPFal9J0E+tl2UQY5g9H16GqhhjYn+39Iei6Q5P4bL4ZubQgTRQTN9nyDc06Z
ezQtGoqZ8kEz/2SyRlkda6PzjSEhgXlc8mCL5J7AW+dMhTHHx2IrosjiCA80kG8=
=c0rK
-----END PGP SIGNATURE-----
Merge tag 'v3.11-rc5' into perf/core
Merge Linux 3.11-rc5, to sync up with the latest upstream fixes since -rc1.
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull nohz improvements from Frederic Weisbecker:
" It mostly contains fixes and full dynticks off-case optimizations. I believe that
distros want to enable this feature so it seems important to optimize the case
where the "nohz_full=" parameter is empty. ie: I'm trying to remove any performance
regression that comes with NO_HZ_FULL=y when the feature is not used.
This patchset improves the current situation a lot (off-case appears to be around 11% faster
with hackbench, although I guess it may vary depending on the configuration but it should be
significantly faster in any case) now there is still some work to do: I can still observe a
remaining loss of 1.6% throughput seen with hackbench compared to CONFIG_NO_HZ_FULL=n. "
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The vtime delta update performed by get_vtime_delta() always check
that the source of the snapshot is valid.
Meanhile the snapshot updaters that rely on get_vtime_delta() also
set the new snapshot origin. But some of them do this right before
the call to get_vtime_delta(), making its debug check useless.
This is easily fixable by moving the snapshot origin update after
the call to get_vtime_delta(). The order doesn't matter there.
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Li Zhong <zhong@linux.vnet.ibm.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Kevin Hilman <khilman@linaro.org>
The cputime accounting in full dynticks can be a subtle
mixup of CPUs using tick based accounting and others using
generic vtime.
As long as the tick can have a share on producing these stats, we
want to scale the result against CFS precise accounting as the tick
can miss some task hiding between the periodic interrupt.
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Li Zhong <zhong@linux.vnet.ibm.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Kevin Hilman <khilman@linaro.org>
If no CPU is in the full dynticks range, we can avoid the full
dynticks cputime accounting through generic vtime along with its
overhead and use the traditional tick based accounting instead.
Let's do this and nope the off case with static keys.
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Li Zhong <zhong@linux.vnet.ibm.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Kevin Hilman <khilman@linaro.org>
get_vtime_delta() must be called under the task vtime_seqlock
with the code that does the cputime accounting flush.
Otherwise the cputime reader can be fooled and run into
a race where it sees the snapshot update but misses the
cputime flush. As a result it can report a cputime that is
way too short.
Fix vtime_account_user() that wasn't complying to that rule.
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Li Zhong <zhong@linux.vnet.ibm.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Kevin Hilman <khilman@linaro.org>
Some generic vtime APIs check if the vtime accounting
is enabled on the local CPU before doing their work.
Some of these are not needed because all their callers already
take care of that. Let's remove the checks on these.
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Li Zhong <zhong@linux.vnet.ibm.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Kevin Hilman <khilman@linaro.org>
Optimize guest entry/exit APIs with static keys. This minimize
the overhead for those who enable CONFIG_NO_HZ_FULL without
always using it. Having no range passed to nohz_full= should
result in the probes overhead to be minimized.
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Li Zhong <zhong@linux.vnet.ibm.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Kevin Hilman <khilman@linaro.org>
Pull scheduler fixes from Ingo Molnar:
"Docbook fixes that make 99% of the diffstat, plus a oneliner fix"
* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched: Ensure update_cfs_shares() is called for parents of continuously-running tasks
sched: Fix some kernel-doc warnings
This is only theoretical, but after try_to_wake_up(p) was changed
to check p->state under p->pi_lock the code like
__set_current_state(TASK_INTERRUPTIBLE);
schedule();
can miss a signal. This is the special case of wait-for-condition,
it relies on try_to_wake_up/schedule interaction and thus it does
not need mb() between __set_current_state() and if(signal_pending).
However, this __set_current_state() can move into the critical
section protected by rq->lock, now that try_to_wake_up() takes
another lock we need to ensure that it can't be reordered with
"if (signal_pending(current))" check inside that section.
The patch is actually one-liner, it simply adds smp_wmb() before
spin_lock_irq(rq->lock). This is what try_to_wake_up() already
does by the same reason.
We turn this wmb() into the new helper, smp_mb__before_spinlock(),
for better documentation and to allow the architectures to change
the default implementation.
While at it, kill smp_mb__after_lock(), it has no callers.
Perhaps we can also add smp_mb__before/after_spinunlock() for
prepare_to_wait().
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Update a stale comment from the old vtime era and document some
locking that might be non obvious.
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Li Zhong <zhong@linux.vnet.ibm.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Kevin Hilman <khilman@linaro.org>
preempt_schedule() and preempt_schedule_context() open
code their preemptability checks.
Use the standard API instead for consolidation.
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Li Zhong <zhong@linux.vnet.ibm.com>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Alex Shi <alex.shi@intel.com>
Cc: Paul Turner <pjt@google.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Vincent Guittot <vincent.guittot@linaro.org>
cgroup is in the process of converting to css (cgroup_subsys_state)
from cgroup as the principal subsystem interface handle. This is
mostly to prepare for the unified hierarchy support where css's will
be created and destroyed dynamically but also helps cleaning up
subsystem implementations as css is usually what they are interested
in anyway.
cgroup_taskset which is used by the subsystem attach methods is the
last cgroup subsystem API which isn't using css as the handle. Update
cgroup_taskset_cur_cgroup() to cgroup_taskset_cur_css() and
cgroup_taskset_for_each() to take @skip_css instead of @skip_cgrp.
The conversions are pretty mechanical. One exception is
cpuset::cgroup_cs(), which lost its last user and got removed.
This patch shouldn't introduce any functional changes.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Matt Helsley <matthltc@us.ibm.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
cgroup is currently in the process of transitioning to using struct
cgroup_subsys_state * as the primary handle instead of struct cgroup.
Please see the previous commit which converts the subsystem methods
for rationale.
This patch converts all cftype file operations to take @css instead of
@cgroup. cftypes for the cgroup core files don't have their subsytem
pointer set. These will automatically use the dummy_css added by the
previous patch and can be converted the same way.
Most subsystem conversions are straight forwards but there are some
interesting ones.
* freezer: update_if_frozen() is also converted to take @css instead
of @cgroup for consistency. This will make the code look simpler
too once iterators are converted to use css.
* memory/vmpressure: mem_cgroup_from_css() needs to be exported to
vmpressure while mem_cgroup_from_cont() can be made static.
Updated accordingly.
* cpu: cgroup_tg() doesn't have any user left. Removed.
* cpuacct: cgroup_ca() doesn't have any user left. Removed.
* hugetlb: hugetlb_cgroup_form_cgroup() doesn't have any user left.
Removed.
* net_cls: cgrp_cls_state() doesn't have any user left. Removed.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
Acked-by: Michal Hocko <mhocko@suse.cz>
Acked-by: Vivek Goyal <vgoyal@redhat.com>
Acked-by: Aristeu Rozanski <aris@redhat.com>
Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: Matt Helsley <matthltc@us.ibm.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Steven Rostedt <rostedt@goodmis.org>
cgroup is currently in the process of transitioning to using struct
cgroup_subsys_state * as the primary handle instead of struct cgroup *
in subsystem implementations for the following reasons.
* With unified hierarchy, subsystems will be dynamically bound and
unbound from cgroups and thus css's (cgroup_subsys_state) may be
created and destroyed dynamically over the lifetime of a cgroup,
which is different from the current state where all css's are
allocated and destroyed together with the associated cgroup. This
in turn means that cgroup_css() should be synchronized and may
return NULL, making it more cumbersome to use.
* Differing levels of per-subsystem granularity in the unified
hierarchy means that the task and descendant iterators should behave
differently depending on the specific subsystem the iteration is
being performed for.
* In majority of the cases, subsystems only care about its part in the
cgroup hierarchy - ie. the hierarchy of css's. Subsystem methods
often obtain the matching css pointer from the cgroup and don't
bother with the cgroup pointer itself. Passing around css fits
much better.
This patch converts all cgroup_subsys methods to take @css instead of
@cgroup. The conversions are mostly straight-forward. A few
noteworthy changes are
* ->css_alloc() now takes css of the parent cgroup rather than the
pointer to the new cgroup as the css for the new cgroup doesn't
exist yet. Knowing the parent css is enough for all the existing
subsystems.
* In kernel/cgroup.c::offline_css(), unnecessary open coded css
dereference is replaced with local variable access.
This patch shouldn't cause any behavior differences.
v2: Unnecessary explicit cgrp->subsys[] deref in css_online() replaced
with local variable @css as suggested by Li Zefan.
Rebased on top of new for-3.12 which includes for-3.11-fixes so
that ->css_free() invocation added by da0a12caff ("cgroup: fix a
leak when percpu_ref_init() fails") is converted too. Suggested
by Li Zefan.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
Acked-by: Michal Hocko <mhocko@suse.cz>
Acked-by: Vivek Goyal <vgoyal@redhat.com>
Acked-by: Aristeu Rozanski <aris@redhat.com>
Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: Matt Helsley <matthltc@us.ibm.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Steven Rostedt <rostedt@goodmis.org>
Currently, controllers have to explicitly follow the cgroup hierarchy
to find the parent of a given css. cgroup is moving towards using
cgroup_subsys_state as the main controller interface construct, so
let's provide a way to climb the hierarchy using just csses.
This patch implements css_parent() which, given a css, returns its
parent. The function is guarnateed to valid non-NULL parent css as
long as the target css is not at the top of the hierarchy.
freezer, cpuset, cpu, cpuacct, hugetlb, memory, net_cls and devices
are converted to use css_parent() instead of accessing cgroup->parent
directly.
* __parent_ca() is dropped from cpuacct and its usage is replaced with
parent_ca(). The only difference between the two was NULL test on
cgroup->parent which is now embedded in css_parent() making the
distinction moot. Note that eventually a css->parent field will be
added to css and the NULL check in css_parent() will go away.
This patch shouldn't cause any behavior differences.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
css (cgroup_subsys_state) is usually embedded in a subsys specific
data structure. Subsystems either use container_of() directly to cast
from css to such data structure or has an accessor function wrapping
such cast. As cgroup as whole is moving towards using css as the main
interface handle, add and update such accessors to ease dealing with
css's.
All accessors explicitly handle NULL input and return NULL in those
cases. While this looks like an extra branch in the code, as all
controllers specific data structures have css as the first field, the
casting doesn't involve any offsetting and the compiler can trivially
optimize out the branch.
* blkio, freezer, cpuset, cpu, cpuacct and net_cls didn't have such
accessor. Added.
* memory, hugetlb and devices already had one but didn't explicitly
handle NULL input. Updated.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
The names of the two struct cgroup_subsys_state accessors -
cgroup_subsys_state() and task_subsys_state() - are somewhat awkward.
The former clashes with the type name and the latter doesn't even
indicate it's somehow related to cgroup.
We're about to revamp large portion of cgroup API, so, let's rename
them so that they're less awkward. Most per-controller usages of the
accessors are localized in accessor wrappers and given the amount of
scheduled changes, this isn't gonna add any noticeable headache.
Rename cgroup_subsys_state() to cgroup_css() and task_subsys_state()
to task_css(). This patch is pure rename.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
Commit 3105b86a9f ("mm: sched: numa: Control enabling and disabling of
NUMA balancing if !SCHED_DEBUG") defined numabalancing_enabled to
control the enabling and disabling of automatic NUMA balancing, but it
is never used.
I believe the intention was to use this in place of sched_feat_numa(NUMA).
Currently, if SCHED_DEBUG is not defined, sched_feat_numa(NUMA) will
never be changed from the initial "false".
Signed-off-by: Dave Kleikamp <dave.kleikamp@oracle.com>
Acked-by: Mel Gorman <mgorman@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
We typically update a task_group's shares within the dequeue/enqueue
path. However, continuously running tasks sharing a CPU are not
subject to these updates as they are only put/picked. Unfortunately,
when we reverted f269ae046 (in 17bc14b7), we lost the augmenting
periodic update that was supposed to account for this; resulting in a
potential loss of fairness.
To fix this, re-introduce the explicit update in
update_cfs_rq_blocked_load() [called via entity_tick()].
Reported-by: Max Hailperin <max@gustavus.edu>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Reviewed-by: Paul Turner <pjt@google.com>
Link: http://lkml.kernel.org/n/tip-9545m3apw5d93ubyrotrj31y@git.kernel.org
Cc: <stable@kernel.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Smart wake-affine is using node-size as the factor currently, but the overhead
of the mask operation is high.
Thus, this patch introduce the 'sd_llc_size' percpu variable, which will record
the highest cache-share domain size, and make it to be the new factor, in order
to reduce the overhead and make it more reasonable.
Tested-by: Davidlohr Bueso <davidlohr.bueso@hp.com>
Tested-by: Michael Wang <wangyun@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Acked-by: Michael Wang <wangyun@linux.vnet.ibm.com>
Cc: Mike Galbraith <efault@gmx.de>
Link: http://lkml.kernel.org/r/51D5008E.6030102@linux.vnet.ibm.com
[ Tidied up the changelog. ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The wake-affine scheduler feature is currently always trying to pull
the wakee close to the waker. In theory this should be beneficial if
the waker's CPU caches hot data for the wakee, and it's also beneficial
in the extreme ping-pong high context switch rate case.
Testing shows it can benefit hackbench up to 15%.
However, the feature is somewhat blind, from which some workloads
such as pgbench suffer. It's also time-consuming algorithmically.
Testing shows it can damage pgbench up to 50% - far more than the
benefit it brings in the best case.
So wake-affine should be smarter and it should realize when to
stop its thankless effort at trying to find a suitable CPU to wake on.
This patch introduces 'wakee_flips', which will be increased each
time the task flips (switches) its wakee target.
So a high 'wakee_flips' value means the task has more than one
wakee, and the bigger the number, the higher the wakeup frequency.
Now when making the decision on whether to pull or not, pay attention to
the wakee with a high 'wakee_flips', pulling such a task may benefit
the wakee. Also imply that the waker will face cruel competition later,
it could be very cruel or very fast depends on the story behind
'wakee_flips', waker therefore suffers.
Furthermore, if waker also has a high 'wakee_flips', that implies that
multiple tasks rely on it, then waker's higher latency will damage all
of them, so pulling wakee seems to be a bad deal.
Thus, when 'waker->wakee_flips / wakee->wakee_flips' becomes
higher and higher, the cost of pulling seems to be worse and worse.
The patch therefore helps the wake-affine feature to stop its pulling
work when:
wakee->wakee_flips > factor &&
waker->wakee_flips > (factor * wakee->wakee_flips)
The 'factor' here is the number of CPUs in the current CPU's NUMA node,
so a bigger node will lead to more pulling since the trial becomes more
severe.
After applying the patch, pgbench shows up to 40% improvements and no regressions.
Tested with 12 cpu x86 server and tip 3.10.0-rc7.
The percentages in the final column highlight the areas with the biggest wins,
all other areas improved as well:
pgbench base smart
| db_size | clients | tps | | tps |
+---------+---------+-------+ +-------+
| 22 MB | 1 | 10598 | | 10796 |
| 22 MB | 2 | 21257 | | 21336 |
| 22 MB | 4 | 41386 | | 41622 |
| 22 MB | 8 | 51253 | | 57932 |
| 22 MB | 12 | 48570 | | 54000 |
| 22 MB | 16 | 46748 | | 55982 | +19.75%
| 22 MB | 24 | 44346 | | 55847 | +25.93%
| 22 MB | 32 | 43460 | | 54614 | +25.66%
| 7484 MB | 1 | 8951 | | 9193 |
| 7484 MB | 2 | 19233 | | 19240 |
| 7484 MB | 4 | 37239 | | 37302 |
| 7484 MB | 8 | 46087 | | 50018 |
| 7484 MB | 12 | 42054 | | 48763 |
| 7484 MB | 16 | 40765 | | 51633 | +26.66%
| 7484 MB | 24 | 37651 | | 52377 | +39.11%
| 7484 MB | 32 | 37056 | | 51108 | +37.92%
| 15 GB | 1 | 8845 | | 9104 |
| 15 GB | 2 | 19094 | | 19162 |
| 15 GB | 4 | 36979 | | 36983 |
| 15 GB | 8 | 46087 | | 49977 |
| 15 GB | 12 | 41901 | | 48591 |
| 15 GB | 16 | 40147 | | 50651 | +26.16%
| 15 GB | 24 | 37250 | | 52365 | +40.58%
| 15 GB | 32 | 36470 | | 50015 | +37.14%
Signed-off-by: Michael Wang <wangyun@linux.vnet.ibm.com>
Cc: Mike Galbraith <efault@gmx.de>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/51D50057.9000809@linux.vnet.ibm.com
[ Improved the changelog. ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The bad thing about update_h_load(), which computes hierarchical load
factor for task groups, is that it is called for each task group in the
system before every load balancer run, and since rebalance can be
triggered very often, this function can eat really a lot of cpu time if
there are many cpu cgroups in the system.
Although the situation was improved significantly by commit a35b646
('sched, cgroup: Reduce rq->lock hold times for large cgroup
hierarchies'), the problem still can arise under some kinds of loads,
e.g. when cpus are switching from idle to busy and back very frequently.
For instance, when I start 1000 of processes that wake up every
millisecond on my 8 cpus host, 'top' and 'perf top' show:
Cpu(s): 17.8%us, 24.3%sy, 0.0%ni, 57.9%id, 0.0%wa, 0.0%hi, 0.0%si
Events: 243K cycles
7.57% [kernel] [k] __schedule
7.08% [kernel] [k] timerqueue_add
6.13% libc-2.12.so [.] usleep
Then if I create 10000 *idle* cpu cgroups (no processes in them), cpu
usage increases significantly although the 'wakers' are still executing
in the root cpu cgroup:
Cpu(s): 19.1%us, 48.7%sy, 0.0%ni, 31.6%id, 0.0%wa, 0.0%hi, 0.7%si
Events: 230K cycles
24.56% [kernel] [k] tg_load_down
5.76% [kernel] [k] __schedule
This happens because this particular kind of load triggers 'new idle'
rebalance very frequently, which requires calling update_h_load(),
which, in turn, calls tg_load_down() for every *idle* cpu cgroup even
though it is absolutely useless, because idle cpu cgroups have no tasks
to pull.
This patch tries to improve the situation by making h_load calculation
proceed only when h_load is really necessary. To achieve this, it
substitutes update_h_load() with update_cfs_rq_h_load(), which computes
h_load only for a given cfs_rq and all its ascendants, and makes the
load balancer call this function whenever it considers if a task should
be pulled, i.e. it moves h_load calculations directly to task_h_load().
For h_load of the same cfs_rq not to be updated multiple times (in case
several tasks in the same cgroup are considered during the same balance
run), the patch keeps the time of the last h_load update for each cfs_rq
and breaks calculation when it finds h_load to be uptodate.
The benefit of it is that h_load is computed only for those cfs_rq's,
which really need it, in particular all idle task groups are skipped.
Although this, in fact, moves h_load calculation under rq lock, it
should not affect latency much, because the amount of work done under rq
lock while trying to pull tasks is limited by sched_nr_migrate.
After the patch applied with the setup described above (1000 wakers in
the root cgroup and 10000 idle cgroups), I get:
Cpu(s): 16.9%us, 24.8%sy, 0.0%ni, 58.4%id, 0.0%wa, 0.0%hi, 0.0%si
Events: 242K cycles
7.57% [kernel] [k] __schedule
6.70% [kernel] [k] timerqueue_add
5.93% libc-2.12.so [.] usleep
Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1373896159-1278-1-git-send-email-vdavydov@parallels.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
cfs_rq is declared twice, fix it.
Also use 'se' instead of '&p->se'.
Signed-off-by: Kirill Tkhai <tkhai@yandex.ru>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Link: http://lkml.kernel.org/r/169201374366727@web6d.yandex.ru
Signed-off-by: Ingo Molnar <mingo@kernel.org>
