The cpuset fields that manage partition root state do not strictly
follow the cpuset locking rule that update to cpuset has to be done
with both the callback_lock and cpuset_mutex held. This is now fixed
by making sure that the locking rule is upheld.
Fixes: 3881b86128 ("cpuset: Add an error state to cpuset.sched.partition")
Fixes: 4b842da276 ("cpuset: Make CPU hotplug work with partition")
Signed-off-by: Waiman Long <longman@redhat.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
This patch adds skb's frag page recycling support based on
the frag page support in page pool.
The performance improves above 10~20% for single thread iperf
TCP flow with IOMMU disabled when iperf server and irq/NAPI
have a different CPU.
The performance improves about 135%(14Gbit to 33Gbit) for single
thread iperf TCP flow when IOMMU is in strict mode and iperf
server shares the same cpu with irq/NAPI.
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Currently page pool only support page recycling when there
is only one user of the page, and the split page reusing
implemented in the most driver can not use the page pool as
bing-pong way of reusing requires the multi user support in
page pool.
Those reusing or recycling has below limitations:
1. page from page pool can only be used be one user in order
for the page recycling to happen.
2. Bing-pong way of reusing in most driver does not support
multi desc using different part of the same page in order
to save memory.
So add multi-users support and frag page recycling in page
pool to overcome the above limitation.
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
For 32 bit systems with 64 bit dma, dma_addr[1] is used to
store the upper 32 bit dma addr, those system should be rare
those days.
For normal system, the dma_addr[1] in 'struct page' is not
used, so we can reuse dma_addr[1] for storing frag count,
which means how many frags this page might be splited to.
In order to simplify the page frag support in the page pool,
the PAGE_POOL_DMA_USE_PP_FRAG_COUNT macro is added to indicate
the 32 bit systems with 64 bit dma, and the page frag support
in page pool is disabled for such system.
The newly added page_pool_set_frag_count() is called to reserve
the maximum frag count before any page frag is passed to the
user. The page_pool_atomic_sub_frag_count_return() is called
when user is done with the page frag.
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Currently, page->pp is cleared and set everytime the page
is recycled, which is unnecessary.
So only set the page->pp when the page is added to the page
pool and only clear it when the page is released from the
page pool.
This is also a preparation to support allocating frag page
in page pool.
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Repair kernel-doc notation in a few places to make it conform to
the expected format.
Fixes the following kernel-doc warnings:
flow.c:296: warning: This comment starts with '/**', but isn't a kernel-doc comment. Refer Documentation/doc-guide/kernel-doc.rst
* Parse vlan tag from vlan header.
flow.c:296: warning: missing initial short description on line:
* Parse vlan tag from vlan header.
flow.c:537: warning: No description found for return value of 'key_extract_l3l4'
flow.c:769: warning: No description found for return value of 'key_extract'
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Pravin B Shelar <pshelar@ovn.org>
Cc: dev@openvswitch.org
Link: https://lore.kernel.org/r/20210808190834.23362-1-rdunlap@infradead.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The functions get_online_cpus() and put_online_cpus() have been
deprecated during the CPU hotplug rework. They map directly to
cpus_read_lock() and cpus_read_unlock().
Replace deprecated CPU-hotplug functions with the official version.
The behavior remains unchanged.
Cc: Tejun Heo <tj@kernel.org>
Reviewed-by: Lai Jiangshan <jiangshanlai@gmail.com>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Tejun Heo <tj@kernel.org>
Replace ida_simple_get() with ida_alloc() and ida_simple_remove() with
ida_free(), the latter is more concise and intuitive.
In addition, if ida_alloc() fails, NULL is returned directly. This
eliminates unnecessary initialization of two local variables and an 'if'
judgment.
Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
The current implementation of rtc_time64_to_tm() contains unnecessary
loops, branches and look-up tables. The new one uses an arithmetic-based
algorithm appeared in [1] and is approximately 4.3 times faster (YMMV).
The drawback is that the new code isn't intuitive and contains many 'magic
numbers' (not unusual for this type of algorithm). However, [1] justifies
all those numbers and, given this function's history, the code is unlikely
to need much maintenance, if any at all.
Add a KUnit test case that checks every day in a 160,000 years interval
starting on 1970-01-01 against the expected result. Add a new config
RTC_LIB_KUNIT_TEST symbol to give the option to run this test suite.
[1] Neri, Schneider, "Euclidean Affine Functions and Applications to
Calendar Algorithms". https://arxiv.org/abs/2102.06959
Signed-off-by: Cassio Neri <cassio.neri@gmail.com>
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Link: https://lore.kernel.org/r/20210624201343.85441-1-cassio.neri@gmail.com
This is most likely going to be 2049 for NFS, but some servers might be
configured to export on a non-standard port. Let's show this information
just in case somebody needs it.
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
I don't support changing it right now, but it could be useful
information for clients with multiple network cards.
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Since bc1c56e9bb transport->srcport may by unset, causing
get_srcport() to return 0 when called. Fix this by querying the port
from the underlying socket instead of the transport.
Fixes: bc1c56e9bb (SUNRPC: prevent port reuse on transports which don't request it)
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
XDP is implemented in the bonding driver by transparently delegating
the XDP program loading, removal and xmit operations to the bonding
slave devices. The overall goal of this work is that XDP programs
can be attached to a bond device *without* any further changes (or
awareness) necessary to the program itself, meaning the same XDP
program can be attached to a native device but also a bonding device.
Semantics of XDP_TX when attached to a bond are equivalent in such
setting to the case when a tc/BPF program would be attached to the
bond, meaning transmitting the packet out of the bond itself using one
of the bond's configured xmit methods to select a slave device (rather
than XDP_TX on the slave itself). Handling of XDP_TX to transmit
using the configured bonding mechanism is therefore implemented by
rewriting the BPF program return value in bpf_prog_run_xdp. To avoid
performance impact this check is guarded by a static key, which is
incremented when a XDP program is loaded onto a bond device. This
approach was chosen to avoid changes to drivers implementing XDP. If
the slave device does not match the receive device, then XDP_REDIRECT
is transparently used to perform the redirection in order to have
the network driver release the packet from its RX ring. The bonding
driver hashing functions have been refactored to allow reuse with
xdp_buff's to avoid code duplication.
The motivation for this change is to enable use of bonding (and
802.3ad) in hairpinning L4 load-balancers such as [1] implemented with
XDP and also to transparently support bond devices for projects that
use XDP given most modern NICs have dual port adapters. An alternative
to this approach would be to implement 802.3ad in user-space and
implement the bonding load-balancing in the XDP program itself, but
is rather a cumbersome endeavor in terms of slave device management
(e.g. by watching netlink) and requires separate programs for native
vs bond cases for the orchestrator. A native in-kernel implementation
overcomes these issues and provides more flexibility.
Below are benchmark results done on two machines with 100Gbit
Intel E810 (ice) NIC and with 32-core 3970X on sending machine, and
16-core 3950X on receiving machine. 64 byte packets were sent with
pktgen-dpdk at full rate. Two issues [2, 3] were identified with the
ice driver, so the tests were performed with iommu=off and patch [2]
applied. Additionally the bonding round robin algorithm was modified
to use per-cpu tx counters as high CPU load (50% vs 10%) and high rate
of cache misses were caused by the shared rr_tx_counter (see patch
2/3). The statistics were collected using "sar -n dev -u 1 10". On top
of that, for ice, further work is in progress on improving the XDP_TX
numbers [4].
-----------------------| CPU |--| rxpck/s |--| txpck/s |----
without patch (1 dev):
XDP_DROP: 3.15% 48.6Mpps
XDP_TX: 3.12% 18.3Mpps 18.3Mpps
XDP_DROP (RSS): 9.47% 116.5Mpps
XDP_TX (RSS): 9.67% 25.3Mpps 24.2Mpps
-----------------------
with patch, bond (1 dev):
XDP_DROP: 3.14% 46.7Mpps
XDP_TX: 3.15% 13.9Mpps 13.9Mpps
XDP_DROP (RSS): 10.33% 117.2Mpps
XDP_TX (RSS): 10.64% 25.1Mpps 24.0Mpps
-----------------------
with patch, bond (2 devs):
XDP_DROP: 6.27% 92.7Mpps
XDP_TX: 6.26% 17.6Mpps 17.5Mpps
XDP_DROP (RSS): 11.38% 117.2Mpps
XDP_TX (RSS): 14.30% 28.7Mpps 27.4Mpps
--------------------------------------------------------------
RSS: Receive Side Scaling, e.g. the packets were sent to a range of
destination IPs.
[1]: https://cilium.io/blog/2021/05/20/cilium-110#standalonelb
[2]: https://lore.kernel.org/bpf/20210601113236.42651-1-maciej.fijalkowski@intel.com/T/#t
[3]: https://lore.kernel.org/bpf/CAHn8xckNXci+X_Eb2WMv4uVYjO2331UWB2JLtXr_58z0Av8+8A@mail.gmail.com/
[4]: https://lore.kernel.org/bpf/20210805230046.28715-1-maciej.fijalkowski@intel.com/T/#t
Signed-off-by: Jussi Maki <joamaki@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Jay Vosburgh <j.vosburgh@gmail.com>
Cc: Veaceslav Falico <vfalico@gmail.com>
Cc: Andy Gospodarek <andy@greyhouse.net>
Cc: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Cc: Magnus Karlsson <magnus.karlsson@intel.com>
Link: https://lore.kernel.org/bpf/20210731055738.16820-4-joamaki@gmail.com
This adds the ndo_xdp_get_xmit_slave hook for transforming XDP_TX
into XDP_REDIRECT after BPF program run when the ingress device
is a bond slave.
The dev_xdp_prog_count is exposed so that slave devices can be checked
for loaded XDP programs in order to avoid the situation where both
bond master and slave have programs loaded according to xdp_state.
Signed-off-by: Jussi Maki <joamaki@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Jay Vosburgh <j.vosburgh@gmail.com>
Cc: Veaceslav Falico <vfalico@gmail.com>
Cc: Andy Gospodarek <andy@greyhouse.net>
Link: https://lore.kernel.org/bpf/20210731055738.16820-3-joamaki@gmail.com
The xprtrdma client code currently relies on the task that initiated the
connect to hold the XPRT_LOCK for the duration of the connection
attempt. If the task is woken early, due to some other event, then that
lock could get released early.
Avoid races by using the same mechanism that the socket code uses of
transferring lock ownership to the RDMA connect worker itself. That
frees us to call rpcrdma_xprt_disconnect() directly since we're now
guaranteed exclusion w.r.t. other callers.
Fixes: 4cf44be6f1 ("xprtrdma: Fix recursion into rpcrdma_xprt_disconnect()")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
We really should not call rpc_wake_up_queued_task_set_status() with
xprt->snd_task as an argument unless we are certain that is actually an
rpc_task.
Fixes: 0445f92c5d ("SUNRPC: Fix disconnection races")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
The intention of the layout barrier is to ensure that we do not update
the layout to match an older value than the current expectation. Fix the
test in pnfs_layout_stateid_blocked() to reflect that it is legal for
the seqid of the stateid to match that of the barrier.
Fixes: aa95edf309 ("NFSv4/pnfs: Fix the layout barrier update")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
A zero value for the layout barrier indicates that it has been cleared
(since seqid '0' is an illegal value), so we should always allow it to
be updated.
Fixes: d29b468da4 ("pNFS/NFSv4: Improve rejection of out-of-order layouts")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Since commit 2b28a7bee4 ("fs, nfs: convert
pnfs_layout_hdr.plh_refcount from atomic_t to refcount_t") it has not
been legal to bump a zero refcount, so the code that tries to allow it
if the NFS_LSEG_VALID flag is still set would cause trouble. Luckily,
NFS_LSEG_VALID has its own refcount so we can never hit this bad code
snippet in practice. Remove it to avoid confusion.
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
If NFS_LAYOUT_RETURN_REQUESTED is set, but there is no value set for
the layout plh_return_seq, we can end up in a livelock loop in which
every layout segment retrieved by a new call to layoutget is immediately
invalidated by pnfs_layout_need_return().
To get around this, we should just set plh_return_seq to the current
value of the layout stateid's seqid.
Fixes: d474f96104 ("NFS: Don't return layout segments that are in use")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Clean up.
Now that there is only one registration mode, there is only one
target "post_send" method: frwr_send(). rpcrdma_post_sends() no
longer adds much value, especially since all of its call sites
ignore the return code value except to check if it's non-zero.
Just have them call frwr_send() directly instead.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Unlike xprtrdma_post_send(), this one can be left enabled all the
time, and should almost never fire. But we do want to know about
immediate errors when they happen.
Note that there is already a similar post_linv_err tracepoint.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
In the vast majority of cases, rc=0. Don't record that in the
post_recvs tracepoint. Instead, add a separate tracepoint that can
be left enabled all the time to capture the very rare immediate
errors returned by ib_post_recv().
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Ensure the tear-down completion is awoken only /after/ we've stopped
fiddling with rpcrdma_rep objects in rpcrdma_post_recvs().
Fixes: 15788d1d10 ("xprtrdma: Do not refresh Receive Queue while it is draining")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
ib_post_send() does not disconnect the QP when it returns an
immediate error. Thus, the code that posts LocalInv has to
explicitly disconnect after an immediate error. This is just
like the frwr_send() callers handle it.
If a disconnect isn't done here, the transport deadlocks.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
When merging one bio to request, if they are discard IO and the queue
supports multi-range discard, we need to return ELEVATOR_DISCARD_MERGE
because both block core and related drivers(nvme, virtio-blk) doesn't
handle mixed discard io merge(traditional IO merge together with
discard merge) well.
Fix the issue by returning ELEVATOR_DISCARD_MERGE in this situation,
so both blk-mq and drivers just need to handle multi-range discard.
Reported-by: Oleksandr Natalenko <oleksandr@natalenko.name>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Tested-by: Oleksandr Natalenko <oleksandr@natalenko.name>
Fixes: 2705dfb209 ("block: fix discard request merge")
Link: https://lore.kernel.org/r/20210729034226.1591070-1-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The client can alter the timeout value after each retransmit. Record
the updated timeout value in the trace log.
Suggested-by: Dai Ngo <dai.ngo@oracle.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Recent patches added RPC_TASK_MOVEABLE, XPRT_OFFLINE, and
XPRT_REMOVE. Update the tracepoint display macros to display these
flags properly.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
In some rare failure modes, the server is actually reading the
transport, but then just dropping the requests on the floor.
TCP_USER_TIMEOUT cannot detect that case.
Prevent such a stuck server from pinning client resources
indefinitely by ensuring that certain idempotent requests
(such as NULL) can time out even if the connection is still
operational.
Otherwise rpc_bind_new_program(), gss_destroy_cred(), or
rpc_clnt_test_and_add_xprt() can wait forever.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
