timewait sockets have a complex refcounting logic.
Once we realize it should be similar to established and
syn_recv sockets, we can use sk_nulls_del_node_init_rcu()
and remove inet_twsk_unhash()
In particular, deferred inet_twsk_put() added in commit
13475a30b6 ("tcp: connect() race with timewait reuse")
looks unecessary : When removing a timewait socket from
ehash or bhash, caller must own a reference on the socket
anyway.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Kernel will crash the same if one of the pointer is NULL anyway.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Merge "Allwinner late changes for 4.2" from Maxime Ripard:
Allwinner late changes for 4.2
A bunch of defconfig changes, and some patches to make the Allwinner H3 and
A33 boot properly.
* tag 'sunxi-late-for-4.2' of https://git.kernel.org/pub/scm/linux/kernel/git/mripard/linux:
ARM: sunxi: Enable simplefb in the defconfig
ARM: Remove deprecated symbol from defconfig files
ARM: sunxi: Add Machine support for A33
ARM: sunxi: Introduce Allwinner H3 support
Documentation: sunxi: Update Allwinner SoC documentation
Oleg Nesterov says:
====================
net: pktgen: fix race between pktgen_thread_worker() and kthread_stop()
I am not familiar with this code and I have no idea how to test
these changes, so 2/2 comes as a separate change. 1/2 looks like
the obvious bugfix, and probably candidate for -stable.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
pktgen_thread_worker() doesn't need to wait for kthread_stop(), it
can simply exit. Just pktgen_create_thread() and pg_net_exit() should
do get_task_struct()/put_task_struct(). kthread_stop(dead_thread) is
fine.
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
pktgen_thread_worker() is obviously racy, kthread_stop() can come
between the kthread_should_stop() check and set_current_state().
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Reported-by: Jan Stancek <jstancek@redhat.com>
Reported-by: Marcelo Leitner <mleitner@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
NCM specs are not actually mandating a specific position in the frame for
the NDP (Network Datagram Pointer). However, some Huawei devices will
ignore our aggregates if it is not placed after the datagrams it points
to. Add support for doing just this, in a per-device configurable way.
While at it, update NCM subdrivers, disabling this functionality in all of
them, except in huawei_cdc_ncm where it is enabled instead.
We aren't making any distinction between different Huawei NCM devices,
based on what the vendor driver does. Standard NCM devices are left
unaffected: if they are compliant, they should be always usable, still
stay on the safe side.
This change has been tested and working with a Huawei E3131 device (which
works regardless of NDP position), a Huawei E3531 (also working both
ways) and an E3372 (which mandates NDP to be after indexed datagrams).
V1->V2:
- corrected wrong NDP acronym definition
- fixed possible NULL pointer dereference
- patch cleanup
V2->V3:
- Properly account for the NDP size when writing new packets to SKB
Signed-off-by: Enrico Mioso <mrkiko.rs@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In commit 'c03abd84634d ("net: ethernet: cpsw: don't requests
IRQs we don't use")', common isr is split into tx and rx, but
in rx isr tx interrupt is also disabledi in cpsw_disable_irq().
So tx interrupts are not handled during rx interrupts and rx
napi completion and results in poor tx performance by 40Mbps.
Fixing by disabling only rx interrupt in rx isr.
Cc: Felipe Balbi <balbi@ti.com>
Cc: <stable@vger.kernel.org> # v4.0+
Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use eth_hw_addr_random() instead of calling random_ether_addr().
Here, this change is setting addr_assign_type to NET_ADDR_RANDOM.
The Coccinelle semantic patch that performs this transformation
is as follows:
@@
identifier a,b;
@@
-random_ether_addr(a->b);
+eth_hw_addr_random(a);
Signed-off-by: Vaishali Thakkar <vthakkar1994@gmail.com>
Tested-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
MV88E6320 and MV88E6321 are largely compatible to MV886352,
but are members of a different chip family.
Signed-off-by: Aleksey S. Kazantsev <ioctl@yandex.ru>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Yuchung Cheng says:
====================
tcp: fixes some congestion control corner cases
This patch series fixes corner cases of TCP congestion control.
First issue is to avoid continuing slow start when cwnd reaches ssthresh.
Second issue is incorrectly processing order of congestion state and
cwnd update when entering fast recovery or undoing cwnd.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
The congestion state and cwnd can be updated in the wrong order.
For example, upon receiving a dubious ACK, we incorrectly raise
the cwnd first (tcp_may_raise_cwnd()/tcp_cong_avoid()) because
the state is still Open, then enter recovery state to reduce cwnd.
For another example, if the ACK indicates spurious timeout or
retransmits, we first revert the cwnd reduction and congestion
state back to Open state. But we don't raise the cwnd even though
the ACK does not indicate any congestion.
To fix this problem we should first call tcp_fastretrans_alert() to
process the dubious ACK and update the congestion state, then call
tcp_may_raise_cwnd() that raises cwnd based on the current state.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Nandita Dukkipati <nanditad@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In the original design slow start is only used to raise cwnd
when cwnd is stricly below ssthresh. It makes little sense
to slow start when cwnd == ssthresh: especially
when hystart has set ssthresh in the initial ramp, or after
recovery when cwnd resets to ssthresh. Not doing so will
also help reduce the buffer bloat slightly.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Nandita Dukkipati <nanditad@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add a helper to test the slow start condition in various congestion
control modules and other places. This is to prepare a slight improvement
in policy as to exactly when to slow start.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Nandita Dukkipati <nanditad@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This change makes it so that the call skb_defer_rx_timestamp will first
check for a phydev before going in and manipulating the skb->data and
skb->len values. By doing this we can avoid unnecessary work on network
devices that don't support phydev. As a result we reduce the total
instruction count needed to process this on most devices.
Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
V1 of this patch contains Eric Dumazet's suggestion to move the per
dst RTAX_QUICKACK check into tcp_in_quickack_mode(). Thanks Eric.
I ran some tests and after setting the "ip route change quickack 1"
knob there were still many delayed ACKs sent. This occured
because when icsk_ack.quick=0 the !icsk_ack.pingpong value is
subsequently ignored as tcp_in_quickack_mode() checks both these
values. The condition for a quick ack to trigger requires
that both icsk_ack.quick != 0 and icsk_ack.pingpong=0. Currently
only icsk_ack.pingpong is controlled by the knob. But the
icsk_ack.quick value changes dynamically depending on heuristics.
The crux of the matter is that delayed acks still cannot be entirely
disabled even with the RTAX_QUICKACK per dst knob enabled. This
patch ensures that a quick ack is always sent when the RTAX_QUICKACK
per dst knob is turned on.
The "ip route change quickack 1" knob was recently added to enable
quickacks. It was modeled around the TCP_QUICKACK setsockopt() option.
This issue is that even with "ip route change quickack 1" enabled
we still see delayed ACKs under some conditions. It would be nice
to be able to completely disable delayed ACKs.
Here is an example:
# netstat -s|grep dela
3 delayed acks sent
For all routes enable the knob
# ip route change quickack 1
Generate some traffic across a slow link and we still see the delayed
acks.
# netstat -s|grep dela
106 delayed acks sent
1 delayed acks further delayed because of locked socket
The issue is that both the "ip route change quickack 1" knob and
the TCP_QUICKACK option set the icsk_ack.pingpong variable to 0.
However at the business end in the __tcp_ack_snd_check() routine,
tcp_in_quickack_mode() checks that both icsk_ack.quick != 0
and icsk_ack.pingpong=0 in order to trigger a quickack. As
icsk_ack.quick is determined by heuristics it can be 0. When
that occurs the icsk_ack.pingpong value is ignored and a delayed
ACK is sent regardless.
This patch moves the RTAX_QUICKACK per dst check into the
tcp_in_quickack_mode() routine which ensures that a quickack is
always sent when the quickack knob is enabled for that dst.
Signed-off-by: Jon Maxwell <jmaxwell37@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Put tick_freeze() under RCU_NONIDLE() to prevent RCU from complaining
about suspicious RCU usage in idle by trace_suspend_resume() called
from there.
While at it, fix a comment related to another usage of RCU_NONIDLE()
in enter_freeze_proper().
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
e48453c386 ("block, cgroup: implement policy-specific per-blkcg
data") updated per-blkcg policy data to be dynamically allocated.
When a policy is registered, its policy data aren't created. Instead,
when the policy is activated on a queue, the policy data are allocated
if there are blkg's (blkcg_gq's) which are attached to a given blkcg.
This is buggy. Consider the following scenario.
1. A blkcg is created. No blkg's attached yet.
2. The policy is registered. No policy data is allocated.
3. The policy is activated on a queue. As the above blkcg doesn't
have any blkg's, it won't allocate the matching blkcg_policy_data.
4. An IO is issued from the blkcg and blkg is created and the blkcg
still doesn't have the matching policy data allocated.
With cfq-iosched, this leads to an oops.
It also doesn't free policy data on policy unregistration assuming
that freeing of all policy data on blkcg destruction should take care
of it; however, this also is incorrect.
1. A blkcg has policy data.
2. The policy gets unregistered but the policy data remains.
3. Another policy gets registered on the same slot.
4. Later, the new policy tries to allocate policy data on the previous
blkcg but the slot is already occupied and gets skipped. The
policy ends up operating on the policy data of the previous policy.
There's no reason to manage blkcg_policy_data lazily. The reason we
do lazy allocation of blkg's is that the number of all possible blkg's
is the product of cgroups and block devices which can reach a
surprising level. blkcg_policy_data is contrained by the number of
cgroups and shouldn't be a problem.
This patch makes blkcg_policy_data to be allocated for all existing
blkcg's on policy registration and freed on unregistration and removes
blkcg_policy_data handling from policy [de]activation paths. This
makes that blkcg_policy_data are created and removed with the policy
they belong to and fixes the above described problems.
Signed-off-by: Tejun Heo <tj@kernel.org>
Fixes: e48453c386 ("block, cgroup: implement policy-specific per-blkcg data")
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Arianna Avanzini <avanzini.arianna@gmail.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Add all_blkcgs list goes through blkcg->all_blkcgs_node and is
protected by blkcg_pol_mutex. This will be used to fix
blkcg_policy_data allocation bug.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Arianna Avanzini <avanzini.arianna@gmail.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
An entry in blkcg_policy[] is stable while there are non-bypassing
in-flight IOs on a request_queue which has the policy activated. This
is why most derefs of blkcg_policy[] don't need explicit locking;
however, blkcg_css_alloc() isn't invoked from IO path and thus doesn't
have this protection and may race policies being added and removed.
Fix it by adding explicit blkcg_pol_mutex protection around
blkcg_policy[] iteration in blkcg_css_alloc().
Signed-off-by: Tejun Heo <tj@kernel.org>
Fixes: e48453c386 ("block, cgroup: implement policy-specific per-blkcg data")
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Arianna Avanzini <avanzini.arianna@gmail.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
blkcg_pol_mutex primarily protects the blkcg_policy array. It also
protects cgroup file type [un]registration during policy addition /
removal. This puts blkcg_pol_mutex outside cgroup internal
synchronization and in turn makes it impossible to grab from blkcg's
cgroup methods as that leads to cyclic dependency.
Another problematic dependency arising from this is through cgroup
interface file deactivation. Removing a cftype requires removing all
files of the type which in turn involves draining all on-going
invocations of the file methods. This means that an interface file
implementation can't grab blkcg_pol_mutex as draining can lead to AA
deadlock.
blkcg_reset_stats() is already in this situation. It currently
trylocks blkcg_pol_mutex and then unwinds and retries the whole
operation on failure, which is cumbersome at best. It has a lengthy
comment explaining how cgroup internal synchronization is involved and
expected to be updated but as explained above this doesn't need cgroup
internal locking to deadlock. It's a self-contained AA deadlock.
The described circular dependencies can be easily broken by moving
cftype [un]registration out of blkcg_pol_mutex and protect them with
an outer mutex. This patch introduces blkcg_pol_register_mutex which
wraps entire policy [un]registration including cftype operations and
shrinks blkcg_pol_mutex critical section. This also makes the trylock
dancing in blkcg_reset_stats() unnecessary. Removed.
This patch is necessary for the following blkcg_policy_data allocation
bug fixes.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Vivek Goyal <vgoyal@redhat.com>
Cc: Arianna Avanzini <avanzini.arianna@gmail.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Merge hpfs updates from Mikulas Patocka.
Mainly fstrim support, with some minor other cleanups.
These were actually sent during the merge window, but I wanted to wait
for the FSTRIM compat handling cleanup before applying them. Mikulas
sent that earlier today.
* emailed patches from Mikulas Patocka <mikulas@twibright.com>:
hpfs: hpfs_error: Remove static buffer, use vsprintf extension %pV instead
hpfs: kstrdup() out of memory handling
hpfs: Remove unessary cast
hpfs: add fstrim support
Mark (and unmark) device nodes with the POPULATE flag as appropriate.
This is required to avoid multi probing when using I2C and device
overlays containing a mux.
This patch is also more careful with the release of the adapter device
which caused a deadlock with muxes, and does not break the build
on !OF since the node flag accessors are not defined then.
Signed-off-by: Pantelis Antoniou <pantelis.antoniou@konsulko.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
Current code returns 0 if fails to read clock-frequency DT property,
fix it. Also add checking return value of clk_prepare_enable and
propagate return value of devm_request_irq.
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
Pull Ceph fixes from Sage Weil:
"There is a fix for CephFS and RBD when used within containers/namespaces,
and a fix for the address learning the client is supposed to do when
initially talking to the Ceph cluster.
There are also two patches updating MAINTAINERS. One breaks out the
common Ceph code shared by fs/ceph and drivers/block/rbd.c into a
separate entry with the appropriate maintainers listed. The second
adds a second reference to the github tree where the Ceph client
development takes place (before it is pushed to korg and then to you).
The goal here is to move closer to a situation where Ilya Dryomov or
one of the other maintainers can push things to you if I am
unavailable. Ilya has done most of the work preparing branches for
upstream recently; you should not be surprised to hear from him if I
am trapped in some internet-less wasteland or hit by a bus or
something. In the meantime, we'll work on getting him added to the
kernel web of trust"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client:
MAINTAINERS: add secondary tree for ceph modules
MAINTAINERS: update ceph entries
libceph: treat sockaddr_storage with uninitialized family as blank
libceph: enable ceph in a non-default network namespace
The FITRIM ioctl has the same arguments on 32-bit and 64-bit
architectures, so we can add it to the list of compatible ioctls and
drop it from compat_ioctl method of various filesystems.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Ted Ts'o <tytso@google.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
At present we don't create efficient ebitmaps when importing NetLabel
category bitmaps. This can present a problem when comparing ebitmaps
since ebitmap_cmp() is very strict about these things and considers
these wasteful ebitmaps not equal when compared to their more
efficient counterparts, even if their values are the same. This isn't
likely to cause problems on 64-bit systems due to a bit of luck on
how NetLabel/CIPSO works and the default ebitmap size, but it can be
a problem on 32-bit systems.
This patch fixes this problem by being a bit more intelligent when
importing NetLabel category bitmaps by skipping over empty sections
which should result in a nice, efficient ebitmap.
Cc: stable@vger.kernel.org # 3.17
Signed-off-by: Paul Moore <pmoore@redhat.com>
The firmware class uevent function accessed the "fw_priv->buf" buffer
without the proper locking and testing for NULL. This is an old bug
(looks like it goes back to 2012 and commit 1244691c73: "firmware
loader: introduce firmware_buf"), but for some reason it's triggering
only now in 4.2-rc1.
Shuah Khan is trying to bisect what it is that causes this to trigger
more easily, but in the meantime let's just fix the bug since others are
hitting it too (at least Ingo reports having seen it as well).
Reported-and-tested-by: Shuah Khan <shuahkh@osg.samsung.com>
Acked-by: Ming Lei <ming.lei@canonical.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Reading page fault handler code I've noticed that under right
circumstances kernel would map anonymous pages into file mappings: if
the VMA doesn't have vm_ops->fault() and the VMA wasn't fully populated
on ->mmap(), kernel would handle page fault to not populated pte with
do_anonymous_page().
Let's change page fault handler to use do_anonymous_page() only on
anonymous VMA (->vm_ops == NULL) and make sure that the VMA is not
shared.
For file mappings without vm_ops->fault() or shred VMA without vm_ops,
page fault on pte_none() entry would lead to SIGBUS.
Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Acked-by: Oleg Nesterov <oleg@redhat.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Willy Tarreau <w@1wt.eu>
Cc: stable@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The Ceph kernel code is primarily developed in the github tree, and only
pushed to the korg tree before going to Linus. If Sage is unavailable and
another maintainer needs to push something upstream, pull requests may
originate from the github tree instead of Sage's korg tree.
Signed-off-by: Sage Weil <sage@redhat.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
- The Ceph common code is used by both fs/ceph and drivers/block/rbd.
Add a separate maintainers entry.
- Add Ilya as libceph maintainer and cephfs submaintainer.
- Attribute Documentation/ABI/testing/sysfs-bus-rbd to rbd.
- ceph-devel@vger.kernel.org should be L, not M in rbd entry.
Signed-off-by: Sage Weil <sage@redhat.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
addr_is_blank() should return true if family is neither AF_INET nor
AF_INET6. This is what its counterpart entity_addr_t::is_blank_ip() is
doing and it is the right thing to do: in process_banner() we check if
our address is blank and if it is "learn" it from our peer. As it is,
we never learn our address and always send out a blank one. This goes
way back to ceph.git commit dd732cbfc1c9 ("use sockaddr_storage; and
some ipv6 support groundwork") from 2009.
While at at, do not open-code ipv6_addr_any() and use INADDR_ANY
constant instead of 0.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Sage Weil <sage@redhat.com>
Grab a reference on a network namespace of the 'rbd map' (in case of
rbd) or 'mount' (in case of ceph) process and use that to open sockets
instead of always using init_net and bailing if network namespace is
anything but init_net. Be careful to not share struct ceph_client
instances between different namespaces and don't add any code in the
!CONFIG_NET_NS case.
This is based on a patch from Hong Zhiguo <zhiguohong@tencent.com>.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Sage Weil <sage@redhat.com>
Cayman does not have vce. There were a few places in the
shared cayman/TV code where we were trying to do vce stuff.
v2: remove -ENOENT check
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
