Pravin Shelar mentioned that GSO could potentially generate
wrong TX checksum if skb has fragments that are overwritten
by the user between the checksum computation and transmit.
He suggested to linearize skbs but this extra copy can be
avoided for normal tcp skbs cooked by tcp_sendmsg().
This patch introduces a new SKB_GSO_SHARED_FRAG flag, set
in skb_shinfo(skb)->gso_type if at least one frag can be
modified by the user.
Typical sources of such possible overwrites are {vm}splice(),
sendfile(), and macvtap/tun/virtio_net drivers.
Tested:
$ netperf -H 7.7.8.84
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to
7.7.8.84 () port 0 AF_INET
Recv Send Send
Socket Socket Message Elapsed
Size Size Size Time Throughput
bytes bytes bytes secs. 10^6bits/sec
87380 16384 16384 10.00 3959.52
$ netperf -H 7.7.8.84 -t TCP_SENDFILE
TCP SENDFILE TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 7.7.8.84 ()
port 0 AF_INET
Recv Send Send
Socket Socket Message Elapsed
Size Size Size Time Throughput
bytes bytes bytes secs. 10^6bits/sec
87380 16384 16384 10.00 3216.80
Performance of the SENDFILE is impacted by the extra allocation and
copy, and because we use order-0 pages, while the TCP_STREAM uses
bigger pages.
Reported-by: Pravin Shelar <pshelar@nicira.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
rcu_dereference occurs in update section. Replacement by
rcu_dereference_protected in order to prevent lockdep
complaint.
Found by Linux Driver Verification project (linuxtesting.org)
Signed-off-by: Denis Efremov <yefremov.denis@gmail.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Ethernet vlan header is not on the packet and kept in the skb->vlan_tci
when it comes from lower dev. This patch inserts vlan header in user
buffer during skb copy on user read.
Signed-off-by: Basil Gor <basil.gor@gmail.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There're several reasons that the vectors need to be validated:
- Return error when caller provides vectors whose num is greater than UIO_MAXIOV.
- Linearize part of skb when userspace provides vectors grater than MAX_SKB_FRAGS.
- Return error when userspace provides vectors whose total length may exceed
- MAX_SKB_FRAGS * PAGE_SIZE.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Current the SKBTX_DEV_ZEROCOPY is set unconditionally after
zerocopy_sg_from_iovec(), this would lead NULL pointer when macvtap
fails to build zerocopy skb because destructor_arg was not
initialized. Solve this by set this flag after the skb were built
successfully.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
When get_user_pages_fast() fails to get all requested pages, we could not use
kfree_skb() to free it as it has not been put in the skb fragments. So we need
to call put_page() instead.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
As the skb fragment were pinned/built from user pages, we should
account the page instead of length for truesize.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This patch fixes the offset calculation when building skb:
- offset1 were used as skb data offset not vector offset
- reset offset to zero only when we advance to next vector
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
rcu_assign_pointer(ptr, NULL) can be safely replaced by
RCU_INIT_POINTER(ptr, NULL)
(old rcu_assign_pointer() macro was testing the NULL value and could
omit the smp_wmb(), but this had to be removed because of compiler
warnings)
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
On systems that create and delete lots of dynamic devices the
31bit linux ifindex fails to fit in the 16bit macvtap minor,
resulting in unusable macvtap devices. I have systems running
automated tests that that hit this condition in just a few days.
Use a linux idr allocator to track which mavtap minor numbers
are available and and to track the association between macvtap
minor numbers and macvtap network devices.
Remove the unnecessary unneccessary check to see if the network
device we have found is indeed a macvtap device. With macvtap
specific data structures it is impossible to find any other
kind of networking device.
Increase the macvtap minor range from 65536 to the full 20 bits
that is supported by linux device numbers. It doesn't solve the
original problem but there is no penalty for a larger minor
device range.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Place macvlan_common_newlink at the end of macvtap_newlink because
failing in newlink after registering your network device is not
supported.
Move device_create into a netdevice creation notifier. The network device
notifier is the only hook that is called after the network device has been
registered with the device layer and before register_network_device returns
success.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
To avoid leaking packets in the receive queue. Add a socket destructor
that will run whenever destroy a macvtap socket.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
To see if it is appropriate to enable the macvtap zero copy feature
don't test the lowerdev network device flags. Instead test the
macvtap network device flags which are a direct copy of the lowerdev
flags. This is important because nothing holds a reference to lowerdev
and on a very bad day we lowerdev could be a pointer to stale memory.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There is a small window in macvtap_open between looking up a
networking device and calling macvtap_set_queue in which
macvtap_del_queues called from macvtap_dellink. After
calling macvtap_del_queues it is totally incorrect to
allow macvtap_set_queue to proceed so prevent success by
reporting that all of the available queues are in use.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit d1b08284 use new frag API but would leave f to be used
uninitialized, this patch fix it.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Only 128 bytes is copied, the rest of data is DMA mapped directly from
userspace.
Signed-off-by: Shirley Ma <xma@...ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There's no need for the guest to validate the checksum if it have been
validated by host nics. So this patch introduces a new flag -
VIRTIO_NET_HDR_F_DATA_VALID which is used to bypass the checksum
examing in guest. The backend (tap/macvtap) may set this flag when
met skbs with CHECKSUM_UNNECESSARY to save cpu utilization.
No feature negotiation is needed as old driver just ignore this flag.
Iperf shows 12%-30% performance improvement for UDP traffic. For TCP,
when gro is on no difference as it produces skb with partial
checksum. But when gro is disabled, 20% or even higher improvement
could be measured by netperf.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
'len' is unsigned of type size_t and can't be negative.
Signed-off-by: Nicolas Kaiser <nikai@nikai.net>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add missing __rcu annotations and helpers.
minor : Fix some rcu_dereference() calls in macvtap
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
After recent changes, (percpu stats on vlan/tunnels...), we dont need
anymore per struct netdev_queue tx_bytes/tx_packets/tx_dropped counters.
Only remaining users are ixgbe, sch_teql, gianfar & macvlan :
1) ixgbe can be converted to use existing tx_ring counters.
2) macvlan incremented txq->tx_dropped, it can use the
dev->stats.tx_dropped counter.
3) sch_teql : almost revert ab35cd4b8f (Use net_device internal stats)
Now we have ndo_get_stats64(), use it, even for "unsigned long"
fields (No need to bring back a struct net_device_stats)
4) gianfar adds a stats structure per tx queue to hold
tx_bytes/tx_packets
This removes a lockdep warning (and possible lockup) in rndis gadget,
calling dev_get_stats() from hard IRQ context.
Ref: http://www.spinics.net/lists/netdev/msg149202.html
Reported-by: Neil Jones <neiljay@gmail.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
CC: Jarek Poplawski <jarkao2@gmail.com>
CC: Alexander Duyck <alexander.h.duyck@intel.com>
CC: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
CC: Sandeep Gopalpet <sandeep.kumar@freescale.com>
CC: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Replace skb->csum_start - skb_headroom(skb) with skb_checksum_start_offset().
Note for usb/smsc95xx: skb->data - skb->head == skb_headroom(skb).
Signed-off-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Signed-off-by: David S. Miller <davem@davemloft.net>
Implement multiqueue facility for macvtap driver. The idea is that
a macvtap device can be opened multiple times and the fd's can be
used to register eg, as backend for vhost.
Signed-off-by: Krishna Kumar <krkumar2@in.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Mark Wagner reported OOM symptoms when sending UDP traffic over
a macvtap link to a kvm receiver.
This appears to be caused by the fact that macvtap packet queues
are unlimited in length. This means that if the receiver can't
keep up with the rate of flow, then we will hit OOM. Of course
it gets worse if the OOM killer then decides to kill the receiver.
This patch imposes a cap on the packet queue length, in the same
way as the tuntap driver, using the device TX queue length.
Please note that macvtap currently has no way of giving congestion
notification, that means the software device TX queue cannot be
used and packets will always be dropped once the macvtap driver
queue fills up.
This shouldn't be a great problem for the scenario where macvtap
is used to feed a kvm receiver, as the traffic is most likely
external in origin so congestion notification can't be applied
anyway.
Of course, if anybody decides to complain about guest-to-guest
UDP packet loss down the track, then we may have to revisit this.
Incidentally, this patch also fixes a real memory leak when
macvtap_get_queue fails.
Chris Wright noticed that for this patch to work, we need a
non-zero TX queue length. This patch includes his work to change
the default macvtap TX queue length to 500.
Reported-by: Mark Wagner <mwagner@redhat.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Acked-by: Chris Wright <chrisw@sous-sol.org>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
This adds TUNSETVNETHDRSZ/TUNGETVNETHDRSZ support
to macvtap.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: David S. Miller <davem@davemloft.net>
sk_callback_lock rwlock actually protects sk->sk_sleep pointer, so we
need two atomic operations (and associated dirtying) per incoming
packet.
RCU conversion is pretty much needed :
1) Add a new structure, called "struct socket_wq" to hold all fields
that will need rcu_read_lock() protection (currently: a
wait_queue_head_t and a struct fasync_struct pointer).
[Future patch will add a list anchor for wakeup coalescing]
2) Attach one of such structure to each "struct socket" created in
sock_alloc_inode().
3) Respect RCU grace period when freeing a "struct socket_wq"
4) Change sk_sleep pointer in "struct sock" by sk_wq, pointer to "struct
socket_wq"
5) Change sk_sleep() function to use new sk->sk_wq instead of
sk->sk_sleep
6) Change sk_has_sleeper() to wq_has_sleeper() that must be used inside
a rcu_read_lock() section.
7) Change all sk_has_sleeper() callers to :
- Use rcu_read_lock() instead of read_lock(&sk->sk_callback_lock)
- Use wq_has_sleeper() to eventually wakeup tasks.
- Use rcu_read_unlock() instead of read_unlock(&sk->sk_callback_lock)
8) sock_wake_async() is modified to use rcu protection as well.
9) Exceptions :
macvtap, drivers/net/tun.c, af_unix use integrated "struct socket_wq"
instead of dynamically allocated ones. They dont need rcu freeing.
Some cleanups or followups are probably needed, (possible
sk_callback_lock conversion to a spinlock for example...).
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit aa395145 (net: sk_sleep() helper) missed three files in the
conversion.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Define a new function to return the waitqueue of a "struct sock".
static inline wait_queue_head_t *sk_sleep(struct sock *sk)
{
return sk->sk_sleep;
}
Change all read occurrences of sk_sleep by a call to this function.
Needed for a future RCU conversion. sk_sleep wont be a field directly
available.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
Added flags field to macvtap_queue to enable/disable processing of
virtio_net_hdr via IFF_VNET_HDR. This flag is checked to prepend virtio_net_hdr
in the receive path and process/skip virtio_net_hdr in the send path.
Original patch by Sridhar, further changes by Arnd.
Signed-off-by: Sridhar Samudrala <sri@us.ibm.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
This adds support for passing a macvtap file descriptor into
vhost-net, much like we already do for tun/tap.
Most of the new code is taken from the respective patch
in the tun driver and may get consolidated in the future.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Sridhar Samudrala <sri@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This reworks the change done by the previous patch
in a more complete way.
The original macvtap code has a number of problems
resulting from the use of RCU for protecting the
access to struct macvtap_queue from open files.
This includes
- need for GFP_ATOMIC allocations for skbs
- potential deadlocks when copy_*_user sleeps
- inability to work with vhost-net
Changing the lifetime of macvtap_queue to always
depend on the open file solves all these. The
RCU reference simply moves one step down to
the reference on the macvlan_dev, which we
only need for nonblocking operations.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Sridhar Samudrala <sri@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The RCU usage in the original code was broken because
there are cases where we possibly sleep with rcu_read_lock
held. As a fix, change the macvtap_file_get_queue to
get a reference on the socket and the netdev instead of
taking the full rcu_read_lock.
Also, change macvtap_file_get_queue failure case to
not require a subsequent macvtap_file_put_queue, as
pointed out by Ed Swierk.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Ed Swierk <eswierk@aristanetworks.com>
Cc: Sridhar Samudrala <sri@us.ibm.com>
Acked-by: Sridhar Samudrala <sri@us.ibm.com>
Acked-by: Ed Swierk <eswierk@aristanetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In order to use macvlan with qemu and other tools that require
a tap file descriptor, the macvtap driver adds a small backend
with a character device with the same interface as the tun
driver, with a minimum set of features.
Macvtap interfaces are created in the same way as macvlan
interfaces using ip link, but the netif is just used as a
handle for configuration and accounting, while the data
goes through the chardev. Each macvtap interface has its
own character device, simplifying permission management
significantly over the generic tun/tap driver.
Cc: Patrick McHardy <kaber@trash.net>
Cc: Stephen Hemminger <shemminger@linux-foundation.org>
Cc: David S. Miller" <davem@davemloft.net>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Or Gerlitz <ogerlitz@voltaire.com>
Cc: netdev@vger.kernel.org
Cc: bridge@lists.linux-foundation.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
