We must try harder to get unique (addr, port) pairs when
doing port autoselection for sockets with SO_REUSEADDR
option set.
We achieve this by adding a relaxation parameter to
inet_csk_bind_conflict. When 'relax' parameter is off
we return a conflict whenever the current searched
pair (addr, port) is not unique.
This tries to address the problems reported in patch:
8d238b25b1
Revert "tcp: bind() fix when many ports are bound"
Tests where ran for creating and binding(0) many sockets
on 100 IPs. The results are, on average:
* 60000 sockets, 600 ports / IP:
* 0.210 s, 620 (IP, port) duplicates without patch
* 0.219 s, no duplicates with patch
* 100000 sockets, 1000 ports / IP:
* 0.371 s, 1720 duplicates without patch
* 0.373 s, no duplicates with patch
* 200000 sockets, 2000 ports / IP:
* 0.766 s, 6900 duplicates without patch
* 0.768 s, no duplicates with patch
* 500000 sockets, 5000 ports / IP:
* 2.227 s, 41500 duplicates without patch
* 2.284 s, no duplicates with patch
Signed-off-by: Alex Copot <alex.mihai.c@gmail.com>
Signed-off-by: Daniel Baluta <dbaluta@ixiacom.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Updates some comments to track RFC6298
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: H.K. Jerry Chu <hkchu@google.com>
Cc: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Convert the per-cpu statistics kept for GRE, IPIP, and SIT tunnels
to use 64 bit statistics.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
net/core/rtnetlink.c: In function ‘rtnl_create_link’:
net/core/rtnetlink.c:1645:3: warning: passing argument 2 of ‘ops->get_tx_queues’ from incompatible pointer type [enabled by default]
net/core/rtnetlink.c:1645:3: note: expected ‘const struct nlattr **’ but argument is of type ‘struct nlattr **’
Signed-off-by: David S. Miller <davem@davemloft.net>
neigh_table_init_no_netlink() is only used in net/core/neighbour.c file.
Signed-off-by: Hiroaki SHIMODA <shimoda.hiroaki@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Most machines dont use UDP encapsulation (L2TP)
Adds a static_key so that udp_queue_rcv_skb() doesnt have to perform a
test if L2TP never setup the encap_rcv on a socket.
Idea of this patch came after Simon Horman proposal to add a hook on TCP
as well.
If static_key is not yet enabled, the fast path does a single JMP .
When static_key is enabled, JMP destination is patched to reach the real
encap_type/encap_rcv logic, possibly adding cache misses.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Simon Horman <horms@verge.net.au>
Cc: dev@openvswitch.org
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix spelling and references in rtnetlink.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Change get_tx_queues, drop unsused arg/return value real_tx_queues,
and use return by value (with error) rather than call by reference.
Probably bonding should just change to LLTX and the whole get_tx_queues
API could disappear!
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Implement aggregation algorithm, combining more data into a single
HSI transfer. 4 different traffic categories are supported:
1. TC_PRIO_CONTROL .. TC_PRIO_MAX (CTL)
2. TC_PRIO_INTERACTIVE (VO)
3. TC_PRIO_INTERACTIVE_BULK (VI)
4. TC_PRIO_BESTEFFORT, TC_PRIO_BULK, TC_PRIO_FILLER (BEBK)
Signed-off-by: Dmitry Tarnyagin <dmitry.tarnyagin@stericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Set traffic class for CAIF packets, based on socket
priority, CAIF protocol type, or type of message.
Traffic class mapping for different packet types:
- control: TC_PRIO_CONTROL;
- flow control: TC_PRIO_CONTROL;
- at: TC_PRIO_CONTROL;
- rfm: TC_PRIO_INTERACTIVE_BULK;
- other sockets: equals to socket's TC;
- network data: no change.
Signed-off-by: Dmitry Tarnyagin <dmitry.tarnyagin@stericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As specified in RFC6106, DNSSL option contains one or more domain names
of DNS suffixes. 8-bit identifier of the DNSSL option type as assigned
by the IANA is 31. This option should also be treated as userland.
Signed-off-by: Alexey I. Froloff <raorn@raorn.name>
Signed-off-by: David S. Miller <davem@davemloft.net>
The firmware may decide to switch channels while already beaconing, e.g.
in response to a cfg80211 connect request on a different vif. Add this
event to notify userspace when an AP or GO interface has successfully
migrated to a new channel, so it can update its configuration
accordingly.
Signed-off-by: Thomas Pedersen <c_tpeder@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
In WoWLAN, we only get the triggers when we actually get
to suspend. As a consequence, drivers currently don't
know that the device should enable wakeup. However, the
device_set_wakeup_enable() API is intended to be called
when the wakeup is enabled, not later when needed.
Add a new set_wakeup() call to cfg80211 and mac80211 to
allow drivers to properly call device_set_wakeup_enable.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
mac80211 currently only supports one hardware queue
per AC. This is already problematic for off-channel
uses since if we go off channel while the BE queue
is full and then try to send an off-channel frame
the frame will never go out. This will become worse
when we support multi-channel since then a queue on
one channel might be full, but we have to stop the
software queue for all channels. That is obviously
not desirable.
To address this problem allow drivers to register
more hardware queues, and allow them to map them to
virtual interfaces. When they stop a hardware queue
the corresponding AC software queues on the correct
interfaces will be stopped as well. Additionally,
there's an off-channel queue to solve that problem
and a per-interface after-DTIM beacon queue. This
allows drivers to manage software queues closer to
how the hardware works.
Currently, there's a limit of 16 hardware queues.
This may or may not be sufficient, we can adjust it
as needed.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The queue mapping redesign that I'm planning to do
will break pure injection unless we handle monitor
interfaces explicitly. One possible option would
be to have the driver tell mac80211 about monitor
mode queues etc., but that would duplicate the API
since we already need to have queue assignments
handled per virtual interface.
So in order to solve this, have a virtual monitor
interface that is added whenever all active vifs
are monitors. We could also use the state of one
of the monitor interfaces, but managing that would
be complicated, so allocate separate state.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Basic rates are added with supported rates IE and extended supported
rates IE.
Signed-off-by: Ashok Nagarajan <ashok@cozybit.com>
Signed-off-by: Thomas Pedersen <thomas@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Report Toffset to userspace.
Let userspace select the mesh synchronization method.
Signed-off-by: Marco Porsch <marco.porsch@s2005.tu-chemnitz.de>
Signed-off-by: Pavel Zubarev <pavel.zubarev@gmail.com>
Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
With the plan to change mac80211's queue API to
not map ACs to queues 1:1, it seems necessary to
clarify some APIs that act on ACs rather than on
queues to spell that out explicitly. Do this.
Also verify that the AC number given is valid.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
This field is never set to anything non-zero in
mac80211, so we should be able to remove it.
Unfortunately though, the iwlwifi and iwlegacy
drivers use it for their internal TX status
processing (which shouldn't be using the rate
control API to start with), so add a new field
"status.antenna" for them, at least for now.
In the future, I plan to use the new field to
hold the hardware queue, while the SKB's queue
mapping holds the AC.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Devices that have internal rate control need to be
notified when the bandwidth or SMPS state changes
just like external rate control algorithms get a
notification now.
Add this notification and clarify the change bits
while at it, the HT_CHANGED bit really meant only
bandwidth changed.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The channel type argument to the rate_update()
callback isn't really the correct way to give
the rate control algorithm about the desired
RX bandwidth of the peer.
Remove this argument, and instead update the
STA capabilities with 20/40 appropriately. The
SMPS update done by this callback works in the
same way, so this makes the callback cleaner.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Since some of the HT code pre-dates 802.11n-2009
some names are wrong. The one that bothers me most
is that "HT operation" is called "HT information"
in our code and that causes confusion.
Rename "HT information" to "HT operation" and also
the control_chan field to primary_chan to match
the name used in the spec.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Acked-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
When the control-rate tables are not set up correctly, it makes
little sense to spam the logs, thus change the WARN_ON to WARN_ON_ONCE.
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
To ensure that old user space versions do not accidentally pick up and
try to use the management channel, use a different channel number.
Reported-by: Keith Packard <keithp@keithp.com>
Acked-by: Johan Hedberg <johan.hedberg@intel.com>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo Padovan <gustavo@padovan.org>
Although not specified in 8021Qaz spec, it could be useful to enable drivers
whose HW supports setting a rate limit for an ETS TC. This patch adds this
optional attribute to DCB netlink. To use it, drivers should implement and
register the callbacks ieee_setmaxrate and ieee_getmaxrate. The units are 64
bits long and specified in Kbps to enable usage over both slow and very fast
networks.
Signed-off-by: Amir Vadai <amirv@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The patch removes unused icmp_ioctl() method definition in
include/net/icmp.h.
Signed-off-by: Rami Rosen <ramirose@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
They were error prone due to an embedded goto, and the entire tree has
been converted away from using them.
Signed-off-by: David S. Miller <davem@davemloft.net>
These macros contain a hidden goto, and are thus extremely error
prone and make code hard to audit.
Signed-off-by: David S. Miller <davem@davemloft.net>
If passed 0 as data_length the (parsed < data_length - 1) test will be
true and cause a buffer overflow. In practice we need at least two bytes
for the element length and type so add a test for it to the very
beginning of the function.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo Padovan <gustavo@padovan.org>
The parsed variable is already incremented inside the for-loop so there
no need to increment it again (not to mention that the code was
incrementing it the wrong amount).
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo Padovan <gustavo@padovan.org>
This patch uses the correct flags for checking the HCI_SSP_ENABLED bit.
Without this authentication request was not being initiated.
Signed-off-by: Hemant Gupta <hemant.gupta@stericsson.com>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Pull networking fixes from David Miller:
1) Name string overrun fix in gianfar driver from Joe Perches.
2) VHOST bug fixes from Michael S. Tsirkin and Nadav Har'El
3) Fix dependencies on xt_LOG netfilter module, from Pablo Neira Ayuso.
4) Fix RCU locking in xt_CT, also from Pablo Neira Ayuso.
5) Add a parameter to skb_add_rx_frag() so we can fix the truesize
adjustments in the drivers that use it. The individual drivers
aren't fixed by this commit, but will be dealt with using follow-on
commits. From Eric Dumazet.
6) Add some device IDs to qmi_wwan driver, from Andrew Bird.
7) Fix a potential rcu_read_lock() imbalancein rt6_fill_node(). From
Eric Dumazet.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net:
net: fix a potential rcu_read_lock() imbalance in rt6_fill_node()
net: add a truesize parameter to skb_add_rx_frag()
gianfar: Fix possible overrun and simplify interrupt name field creation
USB: qmi_wwan: Add ZTE (Vodafone) K3570-Z and K3571-Z net interfaces
USB: option: Ignore ZTE (Vodafone) K3570/71 net interfaces
USB: qmi_wwan: Add ZTE (Vodafone) K3565-Z and K4505-Z net interfaces
qlcnic: Bug fix for LRO
netfilter: nf_conntrack: permanently attach timeout policy to conntrack
netfilter: xt_CT: fix assignation of the generic protocol tracker
netfilter: xt_CT: missing rcu_read_lock section in timeout assignment
netfilter: cttimeout: fix dependency with l4protocol conntrack module
netfilter: xt_LOG: use CONFIG_IP6_NF_IPTABLES instead of CONFIG_IPV6
vhost: fix release path lockdep checks
vhost: don't forget to schedule()
tools/virtio: stub out strong barriers
tools/virtio: add linux/hrtimer.h stub
tools/virtio: add linux/module.h stub
The station_info struct had demanded dBm signal values, but the
cfg80211 wireless extensions implementation was also accepting
"unspecified" (i.e. RSSI) unit values while the nl80211 code was
completely unaware of them. Resolve this by formally allowing the
"unspecified" units while making nl80211 ignore them.
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Reviewed-by: Johannes Berg <johannes@sipsolutions.net>
At least on ia64 the (bogus) use of xchg() here results in the compiler
warning about an unused expression result. As only an assignment is
intended here, convert it to such.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
"[RFC PATCH 0/2] audit of linux/device.h users in include/*"
https://lkml.org/lkml/2012/3/4/159
--
Nearly every subsystem has some kind of header with a proto like:
void foo(struct device *dev);
and yet there is no reason for most of these guys to care about the
sub fields within the device struct. This allows us to significantly
reduce the scope of headers including headers. For this instance, a
reduction of about 40% is achieved by replacing the include with the
simple fact that the device is some kind of a struct.
Unlike the much larger module.h cleanup, this one is simply two
commits. One to fix the implicit <linux/device.h> users, and then
one to delete the device.h includes from the linux/include/ dir
wherever possible.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)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=3j4+
-----END PGP SIGNATURE-----
Merge tag 'device-for-3.4' of git://git.kernel.org/pub/scm/linux/kernel/git/paulg/linux
Pull <linux/device.h> avoidance patches from Paul Gortmaker:
"Nearly every subsystem has some kind of header with a proto like:
void foo(struct device *dev);
and yet there is no reason for most of these guys to care about the
sub fields within the device struct. This allows us to significantly
reduce the scope of headers including headers. For this instance, a
reduction of about 40% is achieved by replacing the include with the
simple fact that the device is some kind of a struct.
Unlike the much larger module.h cleanup, this one is simply two
commits. One to fix the implicit <linux/device.h> users, and then one
to delete the device.h includes from the linux/include/ dir wherever
possible."
* tag 'device-for-3.4' of git://git.kernel.org/pub/scm/linux/kernel/git/paulg/linux:
device.h: audit and cleanup users in main include dir
device.h: cleanup users outside of linux/include (C files)
"[RFC - PATCH 0/7] consolidation of BUG support code."
https://lkml.org/lkml/2012/1/26/525
--
The changes shown here are to unify linux's BUG support under
the one <linux/bug.h> file. Due to historical reasons, we have
some BUG code in bug.h and some in kernel.h -- i.e. the support for
BUILD_BUG in linux/kernel.h predates the addition of linux/bug.h,
but old code in kernel.h wasn't moved to bug.h at that time. As
a band-aid, kernel.h was including <asm/bug.h> to pseudo link them.
This has caused confusion[1] and general yuck/WTF[2] reactions.
Here is an example that violates the principle of least surprise:
CC lib/string.o
lib/string.c: In function 'strlcat':
lib/string.c:225:2: error: implicit declaration of function 'BUILD_BUG_ON'
make[2]: *** [lib/string.o] Error 1
$
$ grep linux/bug.h lib/string.c
#include <linux/bug.h>
$
We've included <linux/bug.h> for the BUG infrastructure and yet we
still get a compile fail! [We've not kernel.h for BUILD_BUG_ON.]
Ugh - very confusing for someone who is new to kernel development.
With the above in mind, the goals of this changeset are:
1) find and fix any include/*.h files that were relying on the
implicit presence of BUG code.
2) find and fix any C files that were consuming kernel.h and
hence relying on implicitly getting some/all BUG code.
3) Move the BUG related code living in kernel.h to <linux/bug.h>
4) remove the asm/bug.h from kernel.h to finally break the chain.
During development, the order was more like 3-4, build-test, 1-2.
But to ensure that git history for bisect doesn't get needless
build failures introduced, the commits have been reorderd to fix
the problem areas in advance.
[1] https://lkml.org/lkml/2012/1/3/90
[2] https://lkml.org/lkml/2012/1/17/414
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)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=PYQT
-----END PGP SIGNATURE-----
Merge tag 'bug-for-3.4' of git://git.kernel.org/pub/scm/linux/kernel/git/paulg/linux
Pull <linux/bug.h> cleanup from Paul Gortmaker:
"The changes shown here are to unify linux's BUG support under the one
<linux/bug.h> file. Due to historical reasons, we have some BUG code
in bug.h and some in kernel.h -- i.e. the support for BUILD_BUG in
linux/kernel.h predates the addition of linux/bug.h, but old code in
kernel.h wasn't moved to bug.h at that time. As a band-aid, kernel.h
was including <asm/bug.h> to pseudo link them.
This has caused confusion[1] and general yuck/WTF[2] reactions. Here
is an example that violates the principle of least surprise:
CC lib/string.o
lib/string.c: In function 'strlcat':
lib/string.c:225:2: error: implicit declaration of function 'BUILD_BUG_ON'
make[2]: *** [lib/string.o] Error 1
$
$ grep linux/bug.h lib/string.c
#include <linux/bug.h>
$
We've included <linux/bug.h> for the BUG infrastructure and yet we
still get a compile fail! [We've not kernel.h for BUILD_BUG_ON.] Ugh -
very confusing for someone who is new to kernel development.
With the above in mind, the goals of this changeset are:
1) find and fix any include/*.h files that were relying on the
implicit presence of BUG code.
2) find and fix any C files that were consuming kernel.h and hence
relying on implicitly getting some/all BUG code.
3) Move the BUG related code living in kernel.h to <linux/bug.h>
4) remove the asm/bug.h from kernel.h to finally break the chain.
During development, the order was more like 3-4, build-test, 1-2. But
to ensure that git history for bisect doesn't get needless build
failures introduced, the commits have been reorderd to fix the problem
areas in advance.
[1] https://lkml.org/lkml/2012/1/3/90
[2] https://lkml.org/lkml/2012/1/17/414"
Fix up conflicts (new radeon file, reiserfs header cleanups) as per Paul
and linux-next.
* tag 'bug-for-3.4' of git://git.kernel.org/pub/scm/linux/kernel/git/paulg/linux:
kernel.h: doesn't explicitly use bug.h, so don't include it.
bug: consolidate BUILD_BUG_ON with other bug code
BUG: headers with BUG/BUG_ON etc. need linux/bug.h
bug.h: add include of it to various implicit C users
lib: fix implicit users of kernel.h for TAINT_WARN
spinlock: macroize assert_spin_locked to avoid bug.h dependency
x86: relocate get/set debugreg fcns to include/asm/debugreg.
In some cases the poll() implementation in a driver has to do different
things depending on the events the caller wants to poll for. An example
is when a driver needs to start a DMA engine if the caller polls for
POLLIN, but doesn't want to do that if POLLIN is not requested but instead
only POLLOUT or POLLPRI is requested. This is something that can happen
in the video4linux subsystem among others.
Unfortunately, the current epoll/poll/select implementation doesn't
provide that information reliably. The poll_table_struct does have it: it
has a key field with the event mask. But once a poll() call matches one
or more bits of that mask any following poll() calls are passed a NULL
poll_table pointer.
Also, the eventpoll implementation always left the key field at ~0 instead
of using the requested events mask.
This was changed in eventpoll.c so the key field now contains the actual
events that should be polled for as set by the caller.
The solution to the NULL poll_table pointer is to set the qproc field to
NULL in poll_table once poll() matches the events, not the poll_table
pointer itself. That way drivers can obtain the mask through a new
poll_requested_events inline.
The poll_table_struct can still be NULL since some kernel code calls it
internally (netfs_state_poll() in ./drivers/staging/pohmelfs/netfs.h). In
that case poll_requested_events() returns ~0 (i.e. all events).
Very rarely drivers might want to know whether poll_wait will actually
wait. If another earlier file descriptor in the set already matched the
events the caller wanted to wait for, then the kernel will return from the
select() call without waiting. This might be useful information in order
to avoid doing expensive work.
A new helper function poll_does_not_wait() is added that drivers can use
to detect this situation. This is now used in sock_poll_wait() in
include/net/sock.h. This was the only place in the kernel that needed
this information.
Drivers should no longer access any of the poll_table internals, but use
the poll_requested_events() and poll_does_not_wait() access functions
instead. In order to enforce that the poll_table fields are now prepended
with an underscore and a comment was added warning against using them
directly.
This required a change in unix_dgram_poll() in unix/af_unix.c which used
the key field to get the requested events. It's been replaced by a call
to poll_requested_events().
For qproc it was especially important to change its name since the
behavior of that field changes with this patch since this function pointer
can now be NULL when that wasn't possible in the past.
Any driver accessing the qproc or key fields directly will now fail to compile.
Some notes regarding the correctness of this patch: the driver's poll()
function is called with a 'struct poll_table_struct *wait' argument. This
pointer may or may not be NULL, drivers can never rely on it being one or
the other as that depends on whether or not an earlier file descriptor in
the select()'s fdset matched the requested events.
There are only three things a driver can do with the wait argument:
1) obtain the key field:
events = wait ? wait->key : ~0;
This will still work although it should be replaced with the new
poll_requested_events() function (which does exactly the same).
This will now even work better, since wait is no longer set to NULL
unnecessarily.
2) use the qproc callback. This could be deadly since qproc can now be
NULL. Renaming qproc should prevent this from happening. There are no
kernel drivers that actually access this callback directly, BTW.
3) test whether wait == NULL to determine whether poll would return without
waiting. This is no longer sufficient as the correct test is now
wait == NULL || wait->_qproc == NULL.
However, the worst that can happen here is a slight performance hit in
the case where wait != NULL and wait->_qproc == NULL. In that case the
driver will assume that poll_wait() will actually add the fd to the set
of waiting file descriptors. Of course, poll_wait() will not do that
since it tests for wait->_qproc. This will not break anything, though.
There is only one place in the whole kernel where this happens
(sock_poll_wait() in include/net/sock.h) and that code will be replaced
by a call to poll_does_not_wait() in the next patch.
Note that even if wait->_qproc != NULL drivers cannot rely on poll_wait()
actually waiting. The next file descriptor from the set might match the
event mask and thus any possible waits will never happen.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Reviewed-by: Jonathan Corbet <corbet@lwn.net>
Reviewed-by: Al Viro <viro@zeniv.linux.org.uk>
Cc: Davide Libenzi <davidel@xmailserver.org>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Cc: Mauro Carvalho Chehab <mchehab@infradead.org>
Cc: David Miller <davem@davemloft.net>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Alex Copot