| Commit message (Collapse) | Author | Files | Lines |
|---|
|
We cannot possibly support SECDISCARD, even if all backend devices would
support it: if our peer is currently unreachable, some instance of the
data may obviously still be recoverable.
We did not set discard_granularity at all. We don't really care (yet),
we only pass them on, so for now, set our granularity to one sector.
blkdev_stack_limits() takes care of the rest.
If we decide we cannot support discards,
not only clear the (not user visible) QUEUE_FLAG_DISCARD,
but set both (user visible) discard_granularity and max_discard_sectors
to zero, to avoid confusion with e.g. lsblk -D.
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
When accessing out meta data area on disk, we double check the
plausibility of the requested sector offsets, and are very noisy about
it if they look suspicious.
During initial read of our "superblock", for "external" meta data,
this triggered because the range estimate returned by
drbd_md_last_sector() was still wrong.
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
Suggested by Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
Apparently we now implicitly get definitions for BITS_PER_PAGE and
BITS_PER_PAGE_MASK from the pid_namespace.h
Instead of renaming our defines, I chose to define only if not yet
defined, but to double check the value if already defined.
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
Since kernel 3.3, we can use snprintf-style arguments
to create a workqueue.
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
The effective data generation ID may be interesting for debugging
purposes of scenarios involving diskless states.
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
In a multiple error scenario, we may end up with a "frozen" Primary,
that has no access to any data (no local disk, no replication link).
If we then resume-io, we try to generate a new data generation id,
which will fail if there is no longer a local disk.
Double check for available local data,
which prevents the NULL pointer deref.
If we are diskless, turn the resume-io in this situation
into the first stage of a "force down", by bumping the "effective" data
gen id, which will prevent later attach or connect to the former data
set without first being demoted (deconfigured).
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
The intention is to reduce CPU utilization. Recent measurements
unveiled that the current performance bottleneck is CPU utilization
on the receiving node. The asender thread became CPU limited.
One of the main points is to eliminate the idr_for_each_entry() loop
from the sending acks code path.
One exception in that is sending back ping_acks. These stay
in the ack-receiver thread. Otherwise the logic becomes too
complicated for no added value.
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
This prepares the next patch where the sending on the meta (or
control) socket is moved to a dedicated workqueue.
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
A D_FAILED disk transitions as quickly as possible to
D_DISKLESS. But in the "unresponsive local disk" case,
there remains a time window where a administrative detach command could
find the disk already failed, but some internal meta data IO against the
unresponsive local disk still pending.
In that case, drbd_md_get_buffer() will return NULL.
Don't unconditionally call drbd_md_put_buffer(), or it will cause
refcount imbalance, and prevent any further re-attach on this volume
(until it is deleted and re-created).
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
The recent (not yet released) backport of the extended state broadcasts
to support the "events2" subcommand of drbdsetup had some glitches.
remember_old_state() would first count all connections with a
net_conf != NULL, then allocate a suitable array, then populate that
array with all connections found to have net_conf != NULL.
This races with the state change to C_STANDALONE,
and the NULL assignment there.
remember_new_state() then iterates over said connection array,
assuming that it would be fully populated.
But rcu_lock() just makes sure the thing some pointer points to,
if any, won't go away. It does not make the pointer itself immutable.
In fact there is no need to "filter" connections based on whether or not
they have a currently valid configuration. Just record them always, if
they don't have a config, that's fine, there will be no change then.
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
Don't blame the peer for being unresponsive,
if we did not even ask the question yet.
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
The only way to make DRBD intentionally call panic is to
set a disk timeout, have that trigger, "abort" some request and complete
to upper layers, then have the backend IO subsystem later complete these
requests successfully regardless.
As the attached IO pages have been recycled for other purposes
meanwhile, this will cause unexpected random memory changes.
To prevent corruption, we rather panic in that case.
Make it obvious from stack traces that this was the case by introducing
drbd_panic_after_delayed_completion_of_aborted_request().
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
Even though we really want to get the state information about our bad
disk to the peer as soon as possible, it is useful to first call the
local-io-error handler.
People may chose to hard-reset the box from there.
If that looks and behaves exactly like a "regular node crash", without
bumping the data generation UUIDs on the peer in between, it makes it
easier to deal with.
If you intend to return from the local-io-error handler, then better
return as quickly as possible to avoid triggering other timeouts.
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
If for some reason the primary lost its disk *and* the replication link
before it is able to communicate the disk loss, probably blocked IO,
then later is able to re-establish the connection, the peer needs to
bump its UUIDs just like it does when peer only loses the disk
and is able to communicate this in time.
Otherwise, a later re-attach of the disk on the primary may start a
resync in the "wrong" direction.
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
"forever"
When detaching, we make sure no application IO is in-flight
by internally suspending IO, then trigger the state change,
wait for the result, and finally internally resume IO again.
Once we triggered the stat change to "Failed",
we expect it to change from Failed to Diskless.
(To avoid races, we actually wait for it to leave "Failed").
On an unresponsive local IO backend, this may not happen, ever.
Don't have a "hung" detach block IO "forever", but resume IO
before waiting for the state change to Diskless.
We may well be able to continue IO to and from a healthy peer.
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
(You should not use disk-timeout anyways,
see the man page for why...)
We add incoming requests to the tail of some ring list.
On local completion, requests are removed from that list.
The timer looks only at the head of that ring list,
so is supposed to only see the oldest request.
All protected by a spinlock.
The request object is created with timestamps zeroed out.
The timestamp was only filled in just before the actual submit.
But to actually submit the request, we need to give up the spinlock.
If you are unlucky, there is no older still pending request, the timer
looks at a new request with timestamp still zero (before it even was
submitted), and 0 + timeout is most likely older than "now".
Better assign the timestamp right when we put the
request object on said ring list.
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
GFP_NOWAIT has a value of 0. I.e. functionality not changed.
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
The lc_destroy() function tests whether its argument is NULL and then
returns immediately. Thus the test around the call is not needed.
This issue was detected by using the Coccinelle software.
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: Roland Kammerer <roland.kammerer@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
The status command originates the drbd9 code base. While for now we
keep the status information in /proc/drbd available, this commit
allows the user base to gracefully migrate their monitoring
infrastructure to the new status reporting interface.
In drbd9 no status information is exposed through /proc/drbd.
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
The events2 command originates from drbd-9 development. It features
more information but requires a incompatible change in output
format.
Therefore the previous events command continues to exist, the new
improved events2 command becomes available now.
This prepares the user-base for a later switch to the complete
drbd9 code base.
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
Instead of using a rwlock for synchronizing state changes across
resources, take the request locks of all resources for global state
changes. Use resources_mutex to serialize global state changes.
This means that taking the request lock of a resource is now enough to
prevent changes of that resource. (Previously, a read lock on the
global state lock was needed as well.)
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
Also change the enum values to all-capital letters.
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
There is no need to have these two as inline functions. In addition,
drbd_should_send_out_of_sync() is only used in a single place, anyway.
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
In drbd-8.4 there is always a single connection per resource,
and there is always exactly one peer_device for a device.
peer_device can not be NULL here.
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
- Changed obsoleted 'P' to 'M' entries.
- Removed the user related mailing list.
- Changed git repos to current versions
Signed-off-by: Roland Kammerer <roland.kammerer@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
Just a comment update on not needing queue_lock, and that we aren't
really adding the request to a timeout list for !mq.
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
Use offset_in_page macro instead of (addr & ~PAGE_MASK).
Signed-off-by: Geliang Tang <geliangtang@163.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
This patch fixes the checkpatch.pl error to genhd.c:
ERROR: do not initialise statics to 0 or NULL
Signed-off-by: Wei Tang <tangwei@cmss.chinamobile.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
This patch fixes the checkpatch.pl error to blk-exec.c:
ERROR: do not initialise globals to 0 or NULL
Signed-off-by: Wei Tang <tangwei@cmss.chinamobile.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
Name the cache after the actual name of the struct.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
We only added the request to the request list for the !blk-mq case,
so we should only delete it in that case as well.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
When we fail various metadata related operations in nvme_queue_rq we
need to unmap the data SGL.
Cc: stable@vger.kernel.org
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
We received a bug report recently when DDW (64-bit direct DMA on Power)
is not enabled for NVMe devices. In that case, we fall back to 32-bit
DMA via the IOMMU, which is always done via 4K TCEs (Translation Control
Entries).
The NVMe device driver, though, assumes that the DMA alignment for the
PRP entries will match the device's page size, and that the DMA aligment
matches the kernel's page aligment. On Power, the the IOMMU page size,
as mentioned above, can be 4K, while the device can have a page size of
8K, while the kernel has a page size of 64K. This eventually trips the
BUG_ON in nvme_setup_prps(), as we have a 'dma_len' that is a multiple
of 4K but not 8K (e.g., 0xF000).
In this particular case of page sizes, we clearly want to use the
IOMMU's page size in the driver. And generally, the NVMe driver in this
function should be using the IOMMU's page size for the default device
page size, rather than the kernel's page size. There is not currently an
API to obtain the IOMMU's page size across all architectures and in the
interest of a stop-gap fix to this functional issue, default the NVMe
device page size to 4K, with the intent of adding such an API and
implementation across all architectures in the next merge window.
With the functionally equivalent v3 of this patch, our hardware test
exerciser survives when using 32-bit DMA; without the patch, the kernel
will BUG within a few minutes.
Signed-off-by: Nishanth Aravamudan <nacc at linux.vnet.ibm.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
I got a crash during a "perf top" session that was caused by a race in
__task_pid_nr_ns() :
pid_nr_ns() was inlined, but apparently compiler chose to read
task->pids[type].pid twice, and the pid->level dereference crashed
because we got a NULL pointer at the second read :
if (pid && ns->level <= pid->level) { // CRASH
Just use RCU API properly to solve this race, and not worry about "perf
top" crashing hosts :(
get_task_pid() can benefit from same fix.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
We had seen lots of reports of this kind issue, so add one
warnning in blk-merge, then it can be triggered easily and
avoid to depend on warning/bug from drivers.
Signed-off-by: Ming Lei <ming.lei@canonical.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
Commit bdced438acd83a(block: setup bi_phys_segments after
splitting) introduces function of computing bio->bi_phys_segments
during bio splitting.
Unfortunately both bio->bi_seg_front_size and bio->bi_seg_back_size
arn't computed, so too many physical segments may be obtained
for one request since both the two are used to check if one segment
across two bios can be possible.
This patch fixes the issue by computing the two variables in
blk_bio_segment_split().
Fixes: bdced438acd83a(block: setup bi_phys_segments after splitting)
Reported-by: Michael Ellerman <mpe@ellerman.id.au>
Reported-by: Mark Salter <msalter@redhat.com>
Tested-by: Laurent Dufour <ldufour@linux.vnet.ibm.com>
Tested-by: Mark Salter <msalter@redhat.com>
Signed-off-by: Ming Lei <ming.lei@canonical.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
Inside blk_bio_segment_split(), previous bvec pointer(bvprvp)
always points to the iterator local variable, which is obviously
wrong, so fix it by pointing to the local variable of 'bvprv'.
Fixes: 5014c311baa2b(block: fix bogus compiler warnings in blk-merge.c)
Cc: stable@kernel.org #4.3
Reported-by: Michael Ellerman <mpe@ellerman.id.au>
Reported-by: Mark Salter <msalter@redhat.com>
Tested-by: Laurent Dufour <ldufour@linux.vnet.ibm.com>
Tested-by: Mark Salter <msalter@redhat.com>
Signed-off-by: Ming Lei <ming.lei@canonical.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
When establishing a thin device's discard limits we cannot rely on the
underlying thin-pool device's discard capabilities (which are inherited
from the thin-pool's underlying data device) given that DM thin devices
must provide discard support even when the thin-pool's underlying data
device doesn't support discards.
Users were exposed to this thin device discard limits regression if
their thin-pool's underlying data device does _not_ support discards.
This regression caused all upper-layers that called the
blkdev_issue_discard() interface to not be able to issue discards to
thin devices (because discard_granularity was 0). This regression
wasn't caught earlier because the device-mapper-test-suite's extensive
'thin-provisioning' discard tests are only ever performed against
thin-pool's with data devices that support discards.
Fix is to have thin_io_hints() test the pool's 'discard_enabled' feature
rather than inferring whether or not a thin device's discard support
should be enabled by looking at the thin-pool's discard_granularity.
Fixes: 216076705 ("dm thin: disable discard support for thin devices if pool's is disabled")
Reported-by: Mike Gerber <mike@sprachgewalt.de>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Cc: stable@vger.kernel.org # 4.1+
|
|
|
|
Adjust kmem_cache_alloc_bulk API before we have any real users.
Adjust API to return type 'int' instead of previously type 'bool'. This
is done to allow future extension of the bulk alloc API.
A future extension could be to allow SLUB to stop at a page boundary, when
specified by a flag, and then return the number of objects.
The advantage of this approach, would make it easier to make bulk alloc
run without local IRQs disabled. With an approach of cmpxchg "stealing"
the entire c->freelist or page->freelist. To avoid overshooting we would
stop processing at a slab-page boundary. Else we always end up returning
some objects at the cost of another cmpxchg.
To keep compatible with future users of this API linking against an older
kernel when using the new flag, we need to return the number of allocated
objects with this API change.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Cc: Vladimir Davydov <vdavydov@virtuozzo.com>
Acked-by: Christoph Lameter <cl@linux.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Initial implementation missed support for kmem cgroup support in
kmem_cache_free_bulk() call, add this.
If CONFIG_MEMCG_KMEM is not enabled, the compiler should be smart enough
to not add any asm code.
Incoming bulk free objects can belong to different kmem cgroups, and
object free call can happen at a later point outside memcg context. Thus,
we need to keep the orig kmem_cache, to correctly verify if a memcg object
match against its "root_cache" (s->memcg_params.root_cache).
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
The call slab_pre_alloc_hook() interacts with kmemgc and is not allowed to
be called several times inside the bulk alloc for loop, due to the call to
memcg_kmem_get_cache().
This would result in hitting the VM_BUG_ON in __memcg_kmem_get_cache.
As suggested by Vladimir Davydov, change slab_post_alloc_hook() to be able
to handle an array of objects.
A subtle detail is, loop iterator "i" in slab_post_alloc_hook() must have
same type (size_t) as size argument. This helps the compiler to easier
realize that it can remove the loop, when all debug statements inside loop
evaluates to nothing. Note, this is only an issue because the kernel is
compiled with GCC option: -fno-strict-overflow
In slab_alloc_node() the compiler inlines and optimizes the invocation of
slab_post_alloc_hook(s, flags, 1, &object) by removing the loop and access
object directly.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Reported-by: Vladimir Davydov <vdavydov@virtuozzo.com>
Suggested-by: Vladimir Davydov <vdavydov@virtuozzo.com>
Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
This change focus on improving the speed of object freeing in the
"slowpath" of kmem_cache_free_bulk.
The calls slab_free (fastpath) and __slab_free (slowpath) have been
extended with support for bulk free, which amortize the overhead of
the (locked) cmpxchg_double.
To use the new bulking feature, we build what I call a detached
freelist. The detached freelist takes advantage of three properties:
1) the free function call owns the object that is about to be freed,
thus writing into this memory is synchronization-free.
2) many freelist's can co-exist side-by-side in the same slab-page
each with a separate head pointer.
3) it is the visibility of the head pointer that needs synchronization.
Given these properties, the brilliant part is that the detached
freelist can be constructed without any need for synchronization. The
freelist is constructed directly in the page objects, without any
synchronization needed. The detached freelist is allocated on the
stack of the function call kmem_cache_free_bulk. Thus, the freelist
head pointer is not visible to other CPUs.
All objects in a SLUB freelist must belong to the same slab-page.
Thus, constructing the detached freelist is about matching objects
that belong to the same slab-page. The bulk free array is scanned is
a progressive manor with a limited look-ahead facility.
Kmem debug support is handled in call of slab_free().
Notice kmem_cache_free_bulk no longer need to disable IRQs. This
only slowed down single free bulk with approx 3 cycles.
Performance data:
Benchmarked[1] obj size 256 bytes on CPU i7-4790K @ 4.00GHz
SLUB fastpath single object quick reuse: 47 cycles(tsc) 11.931 ns
To get stable and comparable numbers, the kernel have been booted with
"slab_merge" (this also improve performance for larger bulk sizes).
Performance data, compared against fallback bulking:
bulk - fallback bulk - improvement with this patch
1 - 62 cycles(tsc) 15.662 ns - 49 cycles(tsc) 12.407 ns- improved 21.0%
2 - 55 cycles(tsc) 13.935 ns - 30 cycles(tsc) 7.506 ns - improved 45.5%
3 - 53 cycles(tsc) 13.341 ns - 23 cycles(tsc) 5.865 ns - improved 56.6%
4 - 52 cycles(tsc) 13.081 ns - 20 cycles(tsc) 5.048 ns - improved 61.5%
8 - 50 cycles(tsc) 12.627 ns - 18 cycles(tsc) 4.659 ns - improved 64.0%
16 - 49 cycles(tsc) 12.412 ns - 17 cycles(tsc) 4.495 ns - improved 65.3%
30 - 49 cycles(tsc) 12.484 ns - 18 cycles(tsc) 4.533 ns - improved 63.3%
32 - 50 cycles(tsc) 12.627 ns - 18 cycles(tsc) 4.707 ns - improved 64.0%
34 - 96 cycles(tsc) 24.243 ns - 23 cycles(tsc) 5.976 ns - improved 76.0%
48 - 83 cycles(tsc) 20.818 ns - 21 cycles(tsc) 5.329 ns - improved 74.7%
64 - 74 cycles(tsc) 18.700 ns - 20 cycles(tsc) 5.127 ns - improved 73.0%
128 - 90 cycles(tsc) 22.734 ns - 27 cycles(tsc) 6.833 ns - improved 70.0%
158 - 99 cycles(tsc) 24.776 ns - 30 cycles(tsc) 7.583 ns - improved 69.7%
250 - 104 cycles(tsc) 26.089 ns - 37 cycles(tsc) 9.280 ns - improved 64.4%
Performance data, compared current in-kernel bulking:
bulk - curr in-kernel - improvement with this patch
1 - 46 cycles(tsc) - 49 cycles(tsc) - improved (cycles:-3) -6.5%
2 - 27 cycles(tsc) - 30 cycles(tsc) - improved (cycles:-3) -11.1%
3 - 21 cycles(tsc) - 23 cycles(tsc) - improved (cycles:-2) -9.5%
4 - 18 cycles(tsc) - 20 cycles(tsc) - improved (cycles:-2) -11.1%
8 - 17 cycles(tsc) - 18 cycles(tsc) - improved (cycles:-1) -5.9%
16 - 18 cycles(tsc) - 17 cycles(tsc) - improved (cycles: 1) 5.6%
30 - 18 cycles(tsc) - 18 cycles(tsc) - improved (cycles: 0) 0.0%
32 - 18 cycles(tsc) - 18 cycles(tsc) - improved (cycles: 0) 0.0%
34 - 78 cycles(tsc) - 23 cycles(tsc) - improved (cycles:55) 70.5%
48 - 60 cycles(tsc) - 21 cycles(tsc) - improved (cycles:39) 65.0%
64 - 49 cycles(tsc) - 20 cycles(tsc) - improved (cycles:29) 59.2%
128 - 69 cycles(tsc) - 27 cycles(tsc) - improved (cycles:42) 60.9%
158 - 79 cycles(tsc) - 30 cycles(tsc) - improved (cycles:49) 62.0%
250 - 86 cycles(tsc) - 37 cycles(tsc) - improved (cycles:49) 57.0%
Performance with normal SLUB merging is significantly slower for
larger bulking. This is believed to (primarily) be an effect of not
having to share the per-CPU data-structures, as tuning per-CPU size
can achieve similar performance.
bulk - slab_nomerge - normal SLUB merge
1 - 49 cycles(tsc) - 49 cycles(tsc) - merge slower with cycles:0
2 - 30 cycles(tsc) - 30 cycles(tsc) - merge slower with cycles:0
3 - 23 cycles(tsc) - 23 cycles(tsc) - merge slower with cycles:0
4 - 20 cycles(tsc) - 20 cycles(tsc) - merge slower with cycles:0
8 - 18 cycles(tsc) - 18 cycles(tsc) - merge slower with cycles:0
16 - 17 cycles(tsc) - 17 cycles(tsc) - merge slower with cycles:0
30 - 18 cycles(tsc) - 23 cycles(tsc) - merge slower with cycles:5
32 - 18 cycles(tsc) - 22 cycles(tsc) - merge slower with cycles:4
34 - 23 cycles(tsc) - 22 cycles(tsc) - merge slower with cycles:-1
48 - 21 cycles(tsc) - 22 cycles(tsc) - merge slower with cycles:1
64 - 20 cycles(tsc) - 48 cycles(tsc) - merge slower with cycles:28
128 - 27 cycles(tsc) - 57 cycles(tsc) - merge slower with cycles:30
158 - 30 cycles(tsc) - 59 cycles(tsc) - merge slower with cycles:29
250 - 37 cycles(tsc) - 56 cycles(tsc) - merge slower with cycles:19
Joint work with Alexander Duyck.
[1] https://github.com/netoptimizer/prototype-kernel/blob/master/kernel/mm/slab_bulk_test01.c
[akpm@linux-foundation.org: BUG_ON -> WARN_ON;return]
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>
Acked-by: Christoph Lameter <cl@linux.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Make it possible to free a freelist with several objects by adjusting API
of slab_free() and __slab_free() to have head, tail and an objects counter
(cnt).
Tail being NULL indicate single object free of head object. This allow
compiler inline constant propagation in slab_free() and
slab_free_freelist_hook() to avoid adding any overhead in case of single
object free.
This allows a freelist with several objects (all within the same
slab-page) to be free'ed using a single locked cmpxchg_double in
__slab_free() and with an unlocked cmpxchg_double in slab_free().
Object debugging on the free path is also extended to handle these
freelists. When CONFIG_SLUB_DEBUG is enabled it will also detect if
objects don't belong to the same slab-page.
These changes are needed for the next patch to bulk free the detached
freelists it introduces and constructs.
Micro benchmarking showed no performance reduction due to this change,
when debugging is turned off (compiled with CONFIG_SLUB_DEBUG).
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>
Acked-by: Christoph Lameter <cl@linux.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Adjust the linker script and map_pages() to map kernel text and data on
physical 1MB huge/large pages.
Signed-off-by: Helge Deller <deller@gmx.de>
|
|
This patch adds huge page support to allow userspace to allocate huge
pages and to use hugetlbfs filesystem on 32- and 64-bit Linux kernels.
A later patch will add kernel support to map kernel text and data on
huge pages.
The only requirement is, that the kernel needs to be compiled for a
PA8X00 CPU (PA2.0 architecture). Older PA1.X CPUs do not support
variable page sizes. 64bit Kernels are compiled for PA2.0 by default.
Technically on parisc multiple physical huge pages may be needed to
emulate standard 2MB huge pages.
Signed-off-by: Helge Deller <deller@gmx.de>
|
