The problem with introducing checksum offload and gso to tun is they
need to set dev->features to enable GSO and/or checksumming, which is
supposed to be done before register_netdevice(), ie. as part of
TUNSETIFF.
Unfortunately, TUNSETIFF has always just ignored flags it doesn't
understand, so there's no good way of detecting whether the kernel
supports new IFF_ flags.
This patch implements a TUNGETFEATURES ioctl which returns all the valid IFF
flags. It could be extended later to include other features.
Here's an example program which uses it:
#include <linux/if_tun.h>
#include <sys/types.h>
#include <sys/ioctl.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <err.h>
#include <stdio.h>
static struct {
unsigned int flag;
const char *name;
} known_flags[] = {
{ IFF_TUN, "TUN" },
{ IFF_TAP, "TAP" },
{ IFF_NO_PI, "NO_PI" },
{ IFF_ONE_QUEUE, "ONE_QUEUE" },
};
int main()
{
unsigned int features, i;
int netfd = open("/dev/net/tun", O_RDWR);
if (netfd < 0)
err(1, "Opening /dev/net/tun");
if (ioctl(netfd, TUNGETFEATURES, &features) != 0) {
printf("Kernel does not support TUNGETFEATURES, guessing\n");
features = (IFF_TUN|IFF_TAP|IFF_NO_PI|IFF_ONE_QUEUE);
}
printf("Available features are: ");
for (i = 0; i < sizeof(known_flags)/sizeof(known_flags[0]); i++) {
if (features & known_flags[i].flag) {
features &= ~known_flags[i].flag;
printf("%s ", known_flags[i].name);
}
}
if (features)
printf("(UNKNOWN %#x)", features);
printf("\n");
return 0;
}
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Acked-by: Max Krasnyansky <maxk@qualcomm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
ip{,v6}_mroute_{set,get}sockopt() should not matter by optimization but
it would be better not to depend on optimization semantically.
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Same as ip6_mr_init(), make ip_mr_init() return errno if fails.
But do not do error handling in inet_init(), just print a msg.
Signed-off-by: Wang Chen <wangchen@cn.fujitsu.com>
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
If do not do it, we will get following issues:
1. Leaving junks after inet6_init failing halfway.
2. Leaving proc and notifier junks after ipv6 modules unloading.
Signed-off-by: Wang Chen <wangchen@cn.fujitsu.com>
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Outgoing interface is selected by the route decision if unspecified.
Let's prefer routes via interface(s) with the address assigned if we
have multiple routes with same cost.
With help from Naohiro Ooiwa <nooiwa@miraclelinux.com>.
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
- If 0, disable DAD.
- If 1, perform DAD (default).
- If >1, perform DAD and disable IPv6 operation if DAD for MAC-based
link-local address has been failed (RFC4862 5.4.5).
We do not follow RFC4862 by default. Refer to the netdev thread entitled
"Linux IPv6 DAD not full conform to RFC 4862 ?"
http://www.spinics.net/lists/netdev/msg52027.html
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Handle interface property strictly when looking up a route
for the loopback address (RFC4291 2.5.3).
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Check the type of the address when adding a new one on interface.
- the unspecified address (::) is always disallowed (RFC4291 2.5.2)
- the loopback address is disallowed unless the interface is (one of)
loopback (RFC4291 2.5.3).
- multicast addresses are disallowed.
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
There are some places in TCP that select one MIB index to
bump snmp statistics like this:
if (<something>)
NET_INC_STATS_BH(<some_id>);
else if (<something_else>)
NET_INC_STATS_BH(<some_other_id>);
...
else
NET_INC_STATS_BH(<default_id>);
or in a more tricky but still similar way.
On the other hand, this NET_INC_STATS_BH is a camouflaged
increment of percpu variable, which is not that small.
Factoring those cases out de-bloats 235 bytes on non-preemptible
i386 config and drives parts of the code into 80 columns.
add/remove: 0/0 grow/shrink: 0/7 up/down: 0/-235 (-235)
function old new delta
tcp_fastretrans_alert 1437 1424 -13
tcp_dsack_set 137 124 -13
tcp_xmit_retransmit_queue 690 676 -14
tcp_try_undo_recovery 283 265 -18
tcp_sacktag_write_queue 1550 1515 -35
tcp_update_reordering 162 106 -56
tcp_retransmit_timer 990 904 -86
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Implemented ethtool callback functions for configuring receive flow
hashing in the niu driver.
Signed-off-by: Santwona Behera <santwona.behera@sun.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Added new interfaces to ethtool to configure receive network flow
distribution across multiple rx rings using hashing.
Signed-off-by: Santwona Behera <santwona.behera@sun.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Socket options SCTP_GET_PEER_ADDR_OLD, SCTP_GET_PEER_ADDR_NUM_OLD,
SCTP_GET_LOCAL_ADDR_OLD, and SCTP_GET_PEER_LOCAL_ADDR_NUM_OLD
have been replaced by newer versions a since 2005. It's time
to officially deprecate them and schedule them for removal.
Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Convert the sysctl values for icmp ratelimit to use milliseconds instead
of jiffies which is based on kernel configured HZ.
Internal kernel jiffies are not a proper unit for any userspace API.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
synchronize_irq() is superfluous when free_irq() call immediately follows it,
because free_irq() also does a synchronize_irq() call of its own.
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
Clean up config/burst value arch-specific setup.
* bcrvalue only varied by its big-endian bit
* crvalue only varied for certain types of x86-32 chips
This should make fealnx quite a bit more portable, without any behavior
change.
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
DE_UNALIGNED_16 is always being passed a u16 *, no need to have the
wrapper with two casts in it, just call get_unaligned directly.
Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
This patch is the minimal amount of code needed to support
wake-on-lan in platform mode properly (i.e. "ethtool -s eth0 wol g"
is sufficient, no additional magic needed) for me.
This is derived from David Brownells patch
(http://lists.laptop.org/pipermail/devel/2007-April/004691.html).
However I decided to move the hook into pci-acpi.c since the other
two pci hooks also live there and pci and acpi are the only users of
the platform_enable_wakeup-hook.
As a 'side-effect' this also makes wake on usb activity work for me
and I had to disable usb wakeup (which is enabled by default) using
the power/wakeup sysfs functionality ("echo disabled >
${sysfs_path_to_device}/power/wakeup").
(BTW I first thought the 'immediate reboot because of usb wake' effect is
caused by the optical mouse generating a wake event, but it rather
seems to be a problem with a flaky secondary usb host controller,
which sees a connected device where nothing is attached)
Signed-off-by: Tobias Diedrich <ranma+kernel@tdiedrich.de>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
We currently don't signal the kernel we that this device can wake
the system. Call device_init_wakeup() to correct this.
Without this device_can_wakeup and device_may_wakeup will return
incorrect values.
Together with the minimized acpi wakeup patch (6/4 ;)), which will
follow in the next mail, this really makes wake-on-lan work for me
as expected (i.e. "ethtool -s eth0 wol g" is sufficient, no
additional magic needed).
Signed-off-by: Tobias Diedrich <ranma+kernel@tdiedrich.de>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
Clean up the following errors and warnings reported by checkpatch.pl:
+ ERROR: Macros with complex values should be enclosed in parenthesis
+ WARNING: __func__ should be used instead of gcc specific __FUNCTION__
+ WARNING: plain inline is preferred over __inline__
+ WARNING: Use #include <linux/io.h> instead of <asm/io.h>
+ WARNING: Use #include <linux/uaccess.h> instead of <asm/uaccess.h>
The changes were verified with by comparing the "objdump -d 8139too.ko"
output which is exactly the same for the old and new version in case of
config CONFIG_8139TOO=m, CONFIG_8139TOO_PIO=n, CONFIG_8139TOO_TUNE_TWISTER=n,
CONFIG_8139TOO_8129=n, CONFIG_8139_OLD_RX_RESET=n.
Software versions used: gcc 4.2.3, objdump 2.18.0.20080103, on elf32-i386.
Signed-off-by: Márton Németh <nm127@freemail.hu>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
Pass buffer length to rndis_command so that rndis_command can read full
response buffer from device instead of max CONTROL_BUFFER_SIZE bytes.
Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
Fix compile error on sh_eth and remove base address macro.
Signed-off-by: Nobuhiro Iwamatsu <iwamatsu.nobuhiro@renesas.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
The call to e1000_clean_tx_irq in e1000_netpoll can race with the call
to e1000_clean_tx_irq in e1000_clean. With a small bit of tweaking to
to netpoll_send_skb to simulate a system that was under extreme stress,
I was able to reproduce these concurrent calls. This can result in
multiple frees to the skbs on the tx ring buffer.
Dropping this call from e1000_netpoll should be fine since we can rely
on the calls in e1000_clean to do what is needed since napi will poll
the hardware just after calling poll_controller.
Signed-off-by: Andy Gospodarek <andy@greyhouse.net>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
This patch makes igb driver ioport-free.
This corrects behavior in probe function so as not to request ioport
resources as long as they are not really needed.
Signed-off-by: Taku Izumi <izumi.taku@jp.fujitsu.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
This patch makes e1000e driver ioport-free.
This corrects behavior in probe function so as not to request ioport
resources as long as they are not really needed.
Signed-off-by: Taku Izumi <izumi.taku@jp.fujitsu.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
We do not want to prolong the situation much longer that e1000
and e1000e support these devices at the same time. As a result,
take out the bandage that was added for the interim period
and remove all the PCI Express device IDs from e1000.
Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
The commit 77d16f450a ("[IPV6] ROUTE:
Unify RT6_F_xxx and RT6_SELECT_F_xxx flags") intended to pass various
routing lookup hints around RT6_LOOKUP_F_xxx flags, but conversion was
missing for rt6_device_match().
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
There is a missing "!" in a conditional statement which is causing entries to
be skipped when dumping the default IPv6 static label entries. This can be
demonstrated by running the following:
# netlabelctl unlbl add default address:::1 \
label:system_u:object_r:unlabeled_t:s0
# netlabelctl -p unlbl list
... you will notice that the entry for the IPv6 localhost address is not
displayed but does exist (works correctly, causes collisions when attempting
to add duplicate entries, etc.).
Signed-off-by: Paul Moore <paul.moore@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When an SKB cannot be chained to a session, the current code attempts
to "restore" its ip_summed field from lro_mgr->ip_summed. However,
lro_mgr->ip_summed does not hold the original value; in fact, we'd
better not touch skb->ip_summed since it is not modified by the code
in the path leading to a failure to chain it. Also use a cleaer
comment to the describe the ip_summed field of struct net_lro_mgr.
Issue raised by Or Gerlitz <ogerlitz@voltaire.com>
Signed-off-by: Eli Cohen <eli@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>
The problem is that while we work w/o the inet_frags.lock even
read-locked the secret rebuild timer may occur (on another CPU, since
BHs are still disabled in the inet_frag_find) and change the rnd seed
for ipv4/6 fragments.
It was caused by my patch fd9e63544c
([INET]: Omit double hash calculations in xxx_frag_intern) late
in the 2.6.24 kernel, so this should probably be queued to -stable.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
I got a problem when I wanted to check if the kernel supports process
event connector, and It seems there's no way to do this check.
At best I can check if the kernel supports connector or not, by looking
into /proc/net/netlink, or maybe checking the return value of bind() to
see if it's ENOENT.
So it would be useful to add /proc/net/connector to list all supported
connectors:
# cat /proc/net/connector
Name ID
connector 4294967295:4294967295
cn_proc 1:1
w1 3:1
Changelog:
- fix memory leak: s/seq_release/single_release
- use spin_lock_bh instead of spin_lock_irqsave
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Acked-by: Evgeniy Polyakov <johnpol@2ka.mipt.ru>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix some doc comments to match function and attribute names in
net/netlink/attr.c.
Signed-off-by: Julius Volz <juliusv@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
I found another case where we are sending information to userspace
in the wrong HZ scale. This should have been fixed back in 2.5 :-(
This means an ABI change but as it stands there is no way for an application
like ss to get the right value.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Due to the CONFIG_'s the value is anyway not correct in userspace.
Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit d62733c8e4
([SCHED]: Qdisc changes and sch_rr added for multiqueue)
added a NET_SCH_RR option that was unused since the code
went unconditionally into sch_prio.
Reported-by: Robert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Note, in the following patch, 'err' is initialized as:
int err = -ENOBUFS;
Signed-off-by: WANG Cong <wcong@critical-links.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
For n:1 'datagram connections' (eg /dev/log), the unix_dgram_sendmsg
routine implements a form of receiver-imposed flow control by
comparing the length of the receive queue of the 'peer socket' with
the max_ack_backlog value stored in the corresponding sock structure,
either blocking the thread which caused the send-routine to be called
or returning EAGAIN. This routine is used by both SOCK_DGRAM and
SOCK_SEQPACKET sockets. The poll-implementation for these socket types
is datagram_poll from core/datagram.c. A socket is deemed to be
writeable by this routine when the memory presently consumed by
datagrams owned by it is less than the configured socket send buffer
size. This is always wrong for PF_UNIX non-stream sockets connected to
server sockets dealing with (potentially) multiple clients if the
abovementioned receive queue is currently considered to be full.
'poll' will then return, indicating that the socket is writeable, but
a subsequent write result in EAGAIN, effectively causing an (usual)
application to 'poll for writeability by repeated send request with
O_NONBLOCK set' until it has consumed its time quantum.
The change below uses a suitably modified variant of the datagram_poll
routines for both type of PF_UNIX sockets, which tests if the
recv-queue of the peer a socket is connected to is presently
considered to be 'full' as part of the 'is this socket
writeable'-checking code. The socket being polled is additionally
put onto the peer_wait wait queue associated with its peer, because the
unix_dgram_recvmsg routine does a wake up on this queue after a
datagram was received and the 'other wakeup call' is done implicitly
as part of skb destruction, meaning, a process blocked in poll
because of a full peer receive queue could otherwise sleep forever
if no datagram owned by its socket was already sitting on this queue.
Among this change is a small (inline) helper routine named
'unix_recvq_full', which consolidates the actual testing code (in three
different places) into a single location.
Signed-off-by: Rainer Weikusat <rweikusat@mssgmbh.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If an skb has nr_frags set to zero but its frag_list is not empty (as
it can happen if software LRO is enabled), and a previous
tcp_read_sock has consumed the linear part of the skb, then
__skb_splice_bits:
(a) incorrectly reports an error and
(b) forgets to update the offset to account for the linear part
Any of the two problems will cause the subsequent __skb_splice_bits
call (the one that handles the frag_list skbs) to either skip data,
or, if the unadjusted offset is greater then the size of the next skb
in the frag_list, make tcp_splice_read loop forever.
Signed-off-by: Octavian Purdila <opurdila@ixiacom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The tcp_mem array which contains limits on the total amount of memory
used by TCP sockets is calculated based on nr_all_pages. On a 32 bits
x86 system, we should base this on the number of lowmem pages.
Signed-off-by: Miquel van Smoorenburg <miquels@cistron.nl>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch fixes an oops in several failure paths in key allocation. This
Oops occurs when freeing a key that has not been linked yet, so the
key->sdata is not set.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Tomas Winkler <tomas.winkler@intel.com>
Acked-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
clock is already cpu-endian (see le32_to_cpu slightly before), so
le64_to_cpu doesn't make much sense.
Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
This fixes a circular locking dependency in the workqueue handling.
The interface work task uses the mac80211 function
ieee80211_iterate_active_interfaces() which grabs the RTNL lock.
However when the interface is brough down, this happens under the RTNL
lock as well, this causes problems because mac80211 will flush the workqueue
during the ifdown event. This causes mac80211 to wait until the driver has
completed all work which can't finish because it is waiting on the RTNL lock.
This is fixed by moving rt2x00 workqueue tasks on a different workqueue,
this workqueue can be flushed when the ieee80211_hw structure is removed
by the driver (when the driver is unloaded) which does not happen under the
RTNL lock.
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
CC [M] drivers/net/wireless/b43/rfkill.o
drivers/net/wireless/b43/rfkill.c: In function ‘b43_rfkill_soft_toggle’:
drivers/net/wireless/b43/rfkill.c:90: warning: enumeration value ‘RFKILL_STATE_HARD_BLOCKED’ not handled in switch
CC [M] drivers/net/wireless/b43legacy/rfkill.o
drivers/net/wireless/b43legacy/rfkill.c: In function ‘b43legacy_rfkill_soft_toggle’:
drivers/net/wireless/b43legacy/rfkill.c:92: warning: enumeration value ‘RFKILL_STATE_HARD_BLOCKED’ not handled in switch
CC [M] drivers/net/wireless/iwlwifi/iwl-rfkill.o
drivers/net/wireless/iwlwifi/iwl-rfkill.c: In function ‘iwl_rfkill_soft_rf_kill’:
drivers/net/wireless/iwlwifi/iwl-rfkill.c:56: warning: enumeration value ‘RFKILL_STATE_HARD_BLOCKED’ not handled in switch
Also handle RFKILL_STATE_{ON,OFF} -> RFKILL_STATE_{UNBLOCKED,SOFT_BLOCKED}
conversion since I'm already here...
Signed-off-by: John W. Linville <linville@tuxdriver.com>
