platform/upstream/kernel-adaptation-pc.git
15 years agoaf_iucv: sync sk shutdown flag if iucv path is quiesced
Hendrik Brueckner [Tue, 21 Apr 2009 23:26:21 +0000 (23:26 +0000)]
af_iucv: sync sk shutdown flag if iucv path is quiesced

If the af_iucv communication partner quiesces the path to shutdown its
receive direction, provide a quiesce callback implementation to shutdown
the (local) send direction. This ensures that both sides are synchronized.

Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoiucv: provide second per-cpu IUCV command parameter block
Ursula Braun [Tue, 21 Apr 2009 23:26:20 +0000 (23:26 +0000)]
iucv: provide second per-cpu IUCV command parameter block

Some of the IUCV commands can be invoked in interrupt context.
Those commands need a different per-cpu IUCV command parameter block,
otherwise they might overwrite an IUCV command parameter of a not yet
finished IUCV command invocation in process context.

Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years ago8390p: Get rid of init_module/cleanup_module
Mateusz Mandera [Wed, 22 Apr 2009 09:24:28 +0000 (02:24 -0700)]
8390p: Get rid of init_module/cleanup_module

This trivial patch just "gets rid of" init_module and cleanup_module
from drivers/net/8390p.c

Signed-off-by: Mateusz Mandera <ormi.linux@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoe1000/e1000e/igb/ixgb: do not use netif_wake_queue un-necessarily
Jesse Brandeburg [Tue, 21 Apr 2009 18:42:05 +0000 (18:42 +0000)]
e1000/e1000e/igb/ixgb: do not use netif_wake_queue un-necessarily

It was pointed out that the Intel wired ethernet drivers do not need to
wake the tx queue since netif_carrier_on/off will take care of the qdisc
management in order to guarantee the correct handling of the transmit
routine enable state.

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agosa1100_ir: fix build breakage
Alexander Beregalov [Wed, 22 Apr 2009 01:12:11 +0000 (18:12 -0700)]
sa1100_ir: fix build breakage

Fix simple typo. Caused by commit
a1de966682551debbe690672e760487cc555c05f
("irda/sa1100_ir: convert to net_device_ops").

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agotun: fix tun_chr_aio_write so that aio works
Michael S. Tsirkin [Mon, 20 Apr 2009 01:26:11 +0000 (01:26 +0000)]
tun: fix tun_chr_aio_write so that aio works

aio_write gets const struct iovec * but tun_chr_aio_write casts this to struct
iovec * and modifies the iovec. As a result, attempts to use io_submit
to send packets to a tun device fail with weird errors such as EINVAL.

Since tun is the only user of skb_copy_datagram_from_iovec, we can
fix this simply by changing the later so that it does not
touch the iovec passed to it.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agotun: fix tun_chr_aio_read so that aio works
Michael S. Tsirkin [Mon, 20 Apr 2009 01:25:59 +0000 (01:25 +0000)]
tun: fix tun_chr_aio_read so that aio works

aio_read gets const struct iovec * but tun_chr_aio_read casts this to struct
iovec * and modifies the iovec. As a result, attempts to use io_submit
to get packets from a tun device fail with weird errors such as EINVAL.

Fix by using the new skb_copy_datagram_const_iovec.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agonet: skb_copy_datagram_const_iovec()
Michael S. Tsirkin [Mon, 20 Apr 2009 01:25:46 +0000 (01:25 +0000)]
net: skb_copy_datagram_const_iovec()

There's an skb_copy_datagram_iovec() to copy out of a paged skb,
but it modifies the iovec, and does not support starting
at an offset in the destination. We want both in tun.c, so let's
add the function.

It's a carbon copy of skb_copy_datagram_iovec() with enough changes to
be annoying.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoucc_geth: Move freeing of TX packets to NAPI context
Joakim Tjernlund [Fri, 17 Apr 2009 12:03:48 +0000 (12:03 +0000)]
ucc_geth: Move freeing of TX packets to NAPI context

This will make the system alot more responsive while ping flooding the
ucc_geth ethernet interface.

Also set NAPI weight to 64 as this is a common value.

Signed-off-by: Joakim Tjernlund <Joakim.Tjernlund@transmode.se>
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agortl8139: unify the struct's name
Jianjun Kong [Mon, 20 Apr 2009 04:05:34 +0000 (04:05 +0000)]
rtl8139: unify the struct's name

unify the struct's name of "struct rtl8139_private *np" to "struct rtl8139_private *tp"

most of them like this:
struct rtl8139_private *tp = netdev_priv(dev);

Signed-off-by: Amos Kong <jianjun@zeuux.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agonet: factor out ethtool invocation of vlan/macvlan drivers
Patrick McHardy [Mon, 20 Apr 2009 04:49:28 +0000 (04:49 +0000)]
net: factor out ethtool invocation of vlan/macvlan drivers

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agousb driver for intellon int51x1 based PLC like devolo dlan duo
Peter Holik [Sat, 18 Apr 2009 07:24:21 +0000 (07:24 +0000)]
usb driver for intellon int51x1 based PLC like devolo dlan duo

usb driver for intellon int51x1 based PLC like devolo dlan duo
with improvements suggested by the guys of the mailinglist:
- name and prefix with int51x1 (Florian Fainelli)
- use conversion functions cpu_to_le16 / le16_to_cpu (Oliver Neukum)
- use pskb_may_pull instead of skb->len (Ilpo Järvinen)
- better code in tx_fixup (Ilpo Järvinen)
- use gotos for error handling (Ilpo Järvinen)
- better description (Jon Smirl)

Signed-off-by: Peter Holik <peter@holik.at>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoexport usbnet_get_ethernet_addr from usbnet and fixed cdc_ether.c
Peter Holik [Sat, 18 Apr 2009 07:24:17 +0000 (07:24 +0000)]
export usbnet_get_ethernet_addr from usbnet and fixed cdc_ether.c

because of using the same function get_ethernet_addr as cdc_ether.c
i export usbnet_get_ethernet_addr from usbnet and fixed cdc_ether
(suggested by Oliver Neukum).

Signed-off-by: Peter Holik <peter@holik.at>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agotg3: Update version to 3.99
Matt Carlson [Mon, 20 Apr 2009 07:13:31 +0000 (07:13 +0000)]
tg3: Update version to 3.99

This patch updates the tg3 version to 3.99.

Signed-off-by: Matt Carlson <mcarlson@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agotg3: Restore LAA sooner in shutdown sequence
Matt Carlson [Mon, 20 Apr 2009 06:58:52 +0000 (06:58 +0000)]
tg3: Restore LAA sooner in shutdown sequence

After a shutdown reset, the LAA needs to be restored before posting the
post-reset signature in shared memory.  If the LAA is not restored
before then, the bootcode will assume the factory default MAC address
and WOL will not work with the LAA.

Signed-off-by: Matt Carlson <mcarlson@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agotg3: Limit CLKREQ fix to A[01] of 57780 asic rev
Matt Carlson [Mon, 20 Apr 2009 06:58:27 +0000 (06:58 +0000)]
tg3: Limit CLKREQ fix to A[01] of 57780 asic rev

This patch restricts the CLKREQ bugfix to the A0 and A1 revisions
of 57780 ASIC rev chips.

Signed-off-by: Matt Carlson <mcarlson@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agotg3: Allow 5761 WOL and LED fixes to 5761S too
Matt Carlson [Mon, 20 Apr 2009 06:58:01 +0000 (06:58 +0000)]
tg3: Allow 5761 WOL and LED fixes to 5761S too

The 5761 WOL and LED fixes used the PCI device ID to as the activation
key.  The 5761S requires the same process.

Signed-off-by: Matt Carlson <mcarlson@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agotg3: Prevent send BD corruption
Matt Carlson [Mon, 20 Apr 2009 06:57:41 +0000 (06:57 +0000)]
tg3: Prevent send BD corruption

On rare occasions, send BD corruptions can occur.  This patch
fixes the problem by increasing the L1 entry threshold to 4
milliseconds.

Signed-off-by: Matt Carlson <mcarlson@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agotg3: Handle NVRAM absent cases
Matt Carlson [Mon, 20 Apr 2009 06:57:14 +0000 (06:57 +0000)]
tg3: Handle NVRAM absent cases

Some 57780 ASIC revision parts do not have NVRAM.  Code the driver so
that it is tolerant of this configuration.

Signed-off-by: Matt Carlson <mcarlson@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agotg3: Allow screaming interrupt detection
Matt Carlson [Mon, 20 Apr 2009 06:55:01 +0000 (06:55 +0000)]
tg3: Allow screaming interrupt detection

The tg3 driver's ISR is coded to accept interrupts as its own if the
status block tag does not equal the last tag the driver has seen.  The
last_tag field is updated from tg3_poll.  In a screaming interrupt
situation from another device sharing tg3's IRQ, tg3_poll does not get
a chance to be called, so the last_tag will always be out of sync with
the status block tag.  Consequently, the driver will continually
declare the screaming interrupts as its own, thus thwarting the
screaming interrupt detection logic.

This patch solves the problem by creating a new last_irq_tag member and
recording the status block tag in the ISR.  The ISR then checks the
last_irq_tag for interrupt ownership.

Many thanks to John Marvin for the detailed bug report and analysis and
Michael Chan for the bugfix.

Signed-off-by: Matt Carlson <mcarlson@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Tested-by: John Marvin <jsm@fc.hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoMerge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
David S. Miller [Tue, 21 Apr 2009 08:32:26 +0000 (01:32 -0700)]
Merge branch 'master' of /linux/kernel/git/davem/net-2.6

Conflicts:
net/core/dev.c

15 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wirel...
David S. Miller [Mon, 20 Apr 2009 22:33:46 +0000 (15:33 -0700)]
Merge branch 'master' of git://git./linux/kernel/git/linville/wireless-2.6

15 years agotg3: Fix SEEPROM accesses
Matt Carlson [Mon, 20 Apr 2009 21:52:29 +0000 (14:52 -0700)]
tg3: Fix SEEPROM accesses

The recent NVRAM patches sanitized how the driver deals with NVRAM
data, but they failed to bring the SEEPROM interfaces inline with
the new strategy.  This patch brings the SEEPROM interfaces up to date.
This patch also reverts commit 0d489ffb76de0fe804cf06a9d4d11fa7342d74b9
("tg3: fix big endian MAC address collection failure").

Signed-off-by: Matt Carlson <mcarlson@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Tested-by: Robin Holt <holt@sgi.com>
Tested-by: James Bottomley <james.bottomley@HansenPartnership.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agomac80211: fix basic rate bitmap calculation
Johannes Berg [Sun, 19 Apr 2009 11:22:11 +0000 (13:22 +0200)]
mac80211: fix basic rate bitmap calculation

"mac80211: fix basic rates setting from association response"
introduced a copy/paste error.

Unfortunately, this not just leads to wrong data being passed
to the driver but is remotely exploitable for some hardware or
driver combinations.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Cc: stable@kernel.org [2.6.29]
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agomac80211: fix beacon loss detection after scan
Kalle Valo [Sun, 19 Apr 2009 05:47:19 +0000 (08:47 +0300)]
mac80211: fix beacon loss detection after scan

Currently beacon loss detection triggers after a scan. A probe request
is sent and a message like this is printed to the log:

wlan0: beacon loss from AP 00:12:17:e7:98:de - sending probe request

But in fact there is no beacon loss, the beacons are just not received
because of the ongoing scan. Fix it by updating last_beacon after
the scan has finished.

Reported-by: Jaswinder Singh Rajput <jaswinder@kernel.org>
Signed-off-by: Kalle Valo <kalle.valo@iki.fi>
Acked-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agonl80211: Make nl80211_send_mlme_event() atomic
Jouni Malinen [Sat, 18 Apr 2009 18:53:15 +0000 (21:53 +0300)]
nl80211: Make nl80211_send_mlme_event() atomic

One of the code paths sending deauth/disassoc events ends up calling
this function with rcu_read_lock held, so we must use GFP_ATOMIC in
allocation routines.

Reported-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Jouni Malinen <j@w1.fi>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agoar9170usb: fix hang on resume
Christian Lamparter [Sat, 18 Apr 2009 15:12:18 +0000 (17:12 +0200)]
ar9170usb: fix hang on resume

This patch fixes a hang on resume when the filesystem is not
available and request_firmware blocks.

However, the device does not accept the firmware on resume.
and it will exit with:

> firmware part 1 upload failed (-71).
> device is in a bad state. please reconnect it!

Reported-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Christian Lamparter <chunkeey@web.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agoMAC80211: Remove unused MAC80211_VERBOSE_SPECT_MGMT_DEBUG.
Robert P. J. Day [Sat, 18 Apr 2009 13:33:55 +0000 (09:33 -0400)]
MAC80211: Remove unused MAC80211_VERBOSE_SPECT_MGMT_DEBUG.

Remove this unused Kconfig variable, which Intel apparently once
promised to make use of but never did.

Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Acked-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agomwl8k: fix module re-insertion bug
Joerg Albert [Sat, 18 Apr 2009 00:10:45 +0000 (02:10 +0200)]
mwl8k: fix module re-insertion bug

swap mwl8k_remove and mwl8k_shutdown functions to allow
"rmmod mwl8k; modprobe mwl8k"

Signed-off-by: Joerg Albert <jal2@gmx.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agop54: deactivate broken powersave function
Christian Lamparter [Fri, 17 Apr 2009 13:14:22 +0000 (15:14 +0200)]
p54: deactivate broken powersave function

This patch deactivates powersave in station mode.

It does not work correctly yet, so the code does more harm than good.

Reported-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Christian Lamparter <chunkeey@web.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agoiwl3945-base.c: Add missing space to debug print
Joe Perches [Thu, 16 Apr 2009 23:23:26 +0000 (16:23 -0700)]
iwl3945-base.c: Add missing space to debug print

"not" is not printed without a space after %pM

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agort2x00: Don't free register information on suspend
Ivo van Doorn [Sat, 28 Mar 2009 19:51:58 +0000 (20:51 +0100)]
rt2x00: Don't free register information on suspend

After suspend & resume the rt2x00 devices won't wakeup
anymore due to a broken register information setup.
The most important problem is the release of the EEPROM
buffer which is completely cleared and never read again
after the suspend.

Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agotun: Fix sk_sleep races when attaching/detaching
Herbert Xu [Sun, 19 Apr 2009 22:35:50 +0000 (22:35 +0000)]
tun: Fix sk_sleep races when attaching/detaching

As the sk_sleep wait queue actually lives in tfile, which may be
detached from the tun device, bad things will happen when we use
sk_sleep after detaching.

Since the tun device is the persistent data structure here (when
requested by the user), it makes much more sense to have the wait
queue live there.  There is no reason to have it in tfile at all
since the only time we can wait is if we have a tun attached.
In fact we already have a wait queue in tun_struct, so we might
as well use it.

Reported-by: Eric W. Biederman <ebiederm@xmission.com>
Tested-by: Christian Borntraeger <borntraeger@de.ibm.com>
Tested-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agotun: Only free a netdev when all tun descriptors are closed
Herbert Xu [Sat, 18 Apr 2009 14:15:52 +0000 (14:15 +0000)]
tun: Only free a netdev when all tun descriptors are closed

The commit c70f182940f988448f3c12a209d18b1edc276e33 ("tun: Fix
races between tun_net_close and free_netdev") fixed a race where
an asynchronous deletion of a tun device can hose a poll(2) on
a tun fd attached to that device.

However, this came at the cost of moving the tun wait queue into
the tun file data structure.  The problem with this is that it
imposes restrictions on when and where the tun device can access
the wait queue since the tun file may change at any time due to
detaching and reattaching.

In particular, now that we need to use the wait queue on the
receive path it becomes difficult to properly synchronise this
with the detachment of the tun device.

This patch solves the original race in a different way.  Since
the race is only because the underlying memory gets freed, we
can prevent it simply by ensuring that we don't do that until
all tun descriptors ever attached to the device (even if they
have since be detached because they may still be sitting in poll)
have been closed.

This is done by using reference counting the attached tun file
descriptors.  The refcount in tun->sk has been reappropriated
for this purpose since it was already being used for that, albeit
from the opposite angle.

Note that we no longer zero tfile->tun since tun_get will return
NULL anyway after the refcount on tfile hits zero.  Instead it
represents whether this device has ever been attached to a device.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agosyncookies: remove last_synq_overflow from struct tcp_sock
Florian Westphal [Sun, 19 Apr 2009 09:43:48 +0000 (09:43 +0000)]
syncookies: remove last_synq_overflow from struct tcp_sock

last_synq_overflow eats 4 or 8 bytes in struct tcp_sock, even
though it is only used when a listening sockets syn queue
is full.

We can (ab)use rx_opt.ts_recent_stamp to store the same information;
it is not used otherwise as long as a socket is in listen state.

Move linger2 around to avoid splitting struct mtu_probe
across cacheline boundary on 32 bit arches.

Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoloopback: packet drops accounting
Eric Dumazet [Fri, 17 Apr 2009 22:03:10 +0000 (22:03 +0000)]
loopback: packet drops accounting

We can in some situations drop packets in netif_rx()

loopback driver does not report these (unlikely) drops to its stats,
and incorrectly change packets/bytes counts.

After this patch applied, "ifconfig lo" can reports these drops as in :

# ifconfig lo
lo        Link encap:Local Loopback
          inet addr:127.0.0.1  Mask:255.0.0.0
          UP LOOPBACK RUNNING  MTU:16436  Metric:1
          RX packets:692562900 errors:3228 dropped:3228 overruns:0 frame:0
          TX packets:692562900 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:2865674174 (2.6 GiB)  TX bytes:2865674174 (2.6 GiB)

I initialy chose to reflect those errors only in tx_dropped/tx_errors, but David
convinced me that it was really RX errors, as loopback_xmit() really starts
a RX process. (calling eth_type_trans() for example, that itself pulls the ethernet header)

These errors are accounted in rx_dropped/rx_errors.

Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agonet: Fix GRO for multiple page fragments
Ben Hutchings [Thu, 16 Apr 2009 08:04:20 +0000 (08:04 +0000)]
net: Fix GRO for multiple page fragments

This loop over fragments in napi_fraginfo_skb() was "interesting".

Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/holtmann/bluet...
David S. Miller [Mon, 20 Apr 2009 09:16:44 +0000 (02:16 -0700)]
Merge branch 'master' of git://git./linux/kernel/git/holtmann/bluetooth-2.6

15 years agonet: fix "compatibility" typos
Marcin Slusarz [Sun, 19 Apr 2009 07:24:24 +0000 (07:24 +0000)]
net: fix "compatibility" typos

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agotcp: fix mid-wq adjustment helper
Ilpo Järvinen [Sat, 18 Apr 2009 05:48:48 +0000 (05:48 +0000)]
tcp: fix mid-wq adjustment helper

Just noticed while doing some new work that the recent
mid-wq adjustment logic will misbehave when FACK is not
in use (happens either due sysctl'ed off or auto-detected
reordering) because I forgot the relevant TCPCB tagbit.

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agonet: sch_netem: Fix an inconsistency in ingress netem timestamps.
Jarek Poplawski [Fri, 17 Apr 2009 10:08:49 +0000 (10:08 +0000)]
net: sch_netem: Fix an inconsistency in ingress netem timestamps.

Alex Sidorenko reported:

"while experimenting with 'netem' we have found some strange behaviour. It
seemed that ingress delay as measured by 'ping' command shows up on some
hosts but not on others.

After some investigation I have found that the problem is that skbuff->tstamp
field value depends on whether there are any packet sniffers enabled. That
is:

- if any ptype_all handler is registered, the tstamp field is as expected
- if there are no ptype_all handlers, the tstamp field does not show the delay"

This patch prevents unnecessary update of tstamp in dev_queue_xmit_nit()
on ingress path (with act_mirred) adding a check, so minimal overhead on
the fast path, but only when sniffers etc. are active.

Since netem at ingress seems to logically emulate a network before a host,
tstamp is zeroed to trigger the update and pretend delays are from the
outside.

Reported-by: Alex Sidorenko <alexandre.sidorenko@hp.com>
Tested-by: Alex Sidorenko <alexandre.sidorenko@hp.com>
Signed-off-by: Jarek Poplawski <jarkao2@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoax25: proc uid file misses header
Alan Cox [Fri, 17 Apr 2009 04:19:36 +0000 (04:19 +0000)]
ax25: proc uid file misses header

This has been broken for a while. I happened to catch it testing because one
app "knew" that the top line of the calls data was the policy line and got
confused.

Put the header back.

Signed-off-by: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agocxgb3: Fix EEH final recovery attempt
Divy Le Ray [Fri, 17 Apr 2009 12:21:27 +0000 (12:21 +0000)]
cxgb3: Fix EEH final recovery attempt

EEH attempts to recover up 6 times.
The last attempt leaves all the ports and adapter down.hen
The driver is then unloaded, bringing the adapter down again
unconditionally. The unload will hang.
Check if the adapter is already down before trying to bring it down again.

Signed-off-by: Divy Le Ray <divy@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agocxgb3: Fix potential msi-x vector leak
Divy Le Ray [Fri, 17 Apr 2009 12:21:22 +0000 (12:21 +0000)]
cxgb3: Fix potential msi-x vector leak

Release vectors when a MSI-X allocation fails.

Signed-off-by: Divy Le Ray <divy@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agocxgb3: fix workqueue flush issues
Divy Le Ray [Fri, 17 Apr 2009 12:21:17 +0000 (12:21 +0000)]
cxgb3: fix workqueue flush issues

The fatal error task can be scheduled while processing an offload packet
in NAPI context when the connection handle is bogus. this can race
with the ports being brought down and the cxgb3 workqueue being flushed.
Stop napi processing before flushing the work queue.

The ULP drivers (iSCSI, iWARP) might also schedule a task on keventd_wk
while releasing a connection handle (cxgb3_offload.c::cxgb3_queue_tid_release()).
The driver however does not flush any work on keventd_wq while being unloaded.
This patch also fixes this.

Also call cancel_delayed_work_sync in place of the the deprecated
cancel_rearming_delayed_workqueue.

Signed-off-by: Divy Le Ray <divy@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agocxgb3: fix link fault handling
Divy Le Ray [Fri, 17 Apr 2009 12:21:11 +0000 (12:21 +0000)]
cxgb3: fix link fault handling

Use the existing periodic task to handle link faults.
The link fault interrupt handler is also called in work queue context,
which is wrong and might cause potential deadlocks.

Signed-off-by: Divy Le Ray <divy@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoe1000: init link state correctly
Jesse Brandeburg [Fri, 17 Apr 2009 20:44:36 +0000 (20:44 +0000)]
e1000: init link state correctly

As reported by Andrew Lutomirski <amluto@gmail.com>

All the intel wired ethernet drivers were calling netif_carrier_off
and netif_stop_queue (or variants) before calling register_netdevice

This is incorrect behavior as was pointed out by davem, and causes
ifconfig and friends to report a strange state before first link
after the driver was loaded.

This apparently confused *some* versions of networkmanager.

Andy tested this for e1000e and confirmed it was working for him.

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Reported-by: Andrew Lutomirski <amluto@gmail.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoe1000e: indicate link down at load
Jesse Brandeburg [Fri, 17 Apr 2009 20:44:34 +0000 (20:44 +0000)]
e1000e: indicate link down at load

As reported by Andrew Lutomirski <amluto@gmail.com>

All the intel wired ethernet drivers were calling netif_carrier_off
and netif_stop_queue (or variants) before calling register_netdevice

This is incorrect behavior as was pointed out by davem, and causes
ifconfig and friends to report a strange state before first link
after the driver was loaded, since without a netif_carrier_off, the stack
assumes carrier_on, but before register_netdev, netlink messages are not
sent out telling link state.

This apparently confused *some* versions of networkmanager.

Andy tested this for e1000e and confirmed it was working for him.

see thread: http://marc.info/?l=linux-netdev&m=123946479705636&w=2

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Andy Lutomirski <amluto@gmail.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoigb: fix link down inital state
Jesse Brandeburg [Fri, 17 Apr 2009 20:44:32 +0000 (20:44 +0000)]
igb: fix link down inital state

As reported by Andrew Lutomirski <amluto@gmail.com>

All the intel wired ethernet drivers were calling netif_carrier_off
and netif_stop_queue (or variants) before calling register_netdevice

This is incorrect behavior as was pointed out by davem, and causes
ifconfig and friends to report a strange state before first link
after the driver was loaded, since without a netif_carrier_off, the stack
assumes carrier_on, but before register_netdev, netlink messages are not
sent out telling link state.

This apparently confused *some* versions of networkmanager.

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Reported-by: Andrew Lutomirski <amluto@gmail.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoixgb: fix link down initial state
Jesse Brandeburg [Fri, 17 Apr 2009 20:44:29 +0000 (20:44 +0000)]
ixgb: fix link down initial state

As reported by Andrew Lutomirski <amluto@gmail.com>

All the intel wired ethernet drivers were calling netif_carrier_off
and netif_stop_queue (or variants) before calling register_netdevice

This is incorrect behavior as was pointed out by davem, and causes
ifconfig and friends to report a strange state before first link
after the driver was loaded, since without a netif_carrier_off, the stack
assumes carrier_on, but before register_netdev, netlink messages are not
sent out telling link state.

This apparently confused *some* versions of networkmanager.

in addition this driver appeared to need a netif_start_queue at
the end of open.

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Reported-by: Andrew Lutomirski <amluto@gmail.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoixgbe: fix link down initial state
Jesse Brandeburg [Fri, 17 Apr 2009 20:44:27 +0000 (20:44 +0000)]
ixgbe: fix link down initial state

As reported by Andrew Lutomirski <amluto@gmail.com>

All the intel wired ethernet drivers were calling netif_carrier_off
and netif_stop_queue (or variants) before calling register_netdevice

This is incorrect behavior as was pointed out by davem, and causes
ifconfig and friends to report a strange state before first link
after the driver was loaded, since without a netif_carrier_off, the stack
assumes carrier_on, but before register_netdev, netlink messages are not
sent out telling link state.

This apparently confused *some* versions of networkmanager.

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Reported-by: Andrew Lutomirski <amluto@gmail.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoiseries_veth: fix build breakage
Alexander Beregalov [Mon, 20 Apr 2009 01:42:58 +0000 (18:42 -0700)]
iseries_veth: fix build breakage

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoBluetooth: Add workaround for wrong HCI event in eSCO setup
Marcel Holtmann [Sun, 19 Apr 2009 17:30:03 +0000 (19:30 +0200)]
Bluetooth: Add workaround for wrong HCI event in eSCO setup

The Broadcom chips with 2.1 firmware handle the fallback case to a SCO
link wrongly when setting up eSCO connections.

  < HCI Command: Setup Synchronous Connection (0x01|0x0028) plen 17
      handle 11 voice setting 0x0060
  > HCI Event: Command Status (0x0f) plen 4
      Setup Synchronous Connection (0x01|0x0028) status 0x00 ncmd 1
  > HCI Event: Connect Complete (0x03) plen 11
      status 0x00 handle 1 bdaddr 00:1E:3A:xx:xx:xx type SCO encrypt 0x01

The Link Manager negotiates the fallback to SCO, but then sends out
a Connect Complete event. This is wrong and the Link Manager should
actually send a Synchronous Connection Complete event if the Setup
Synchronous Connection has been used. Only the remote side is allowed
to use Connect Complete to indicate the missing support for eSCO in
the host stack.

This patch adds a workaround for this which clearly should not be
needed, but reality is that broken Broadcom devices are deployed.

Based on a report by Ville Tervo <ville.tervo@nokia.com>

Signed-off-by: Marcel Holtman <marcel@holtmann.org>
15 years agoBluetooth: Fallback from eSCO to SCO on unspecified error
Marcel Holtmann [Sun, 19 Apr 2009 17:14:14 +0000 (19:14 +0200)]
Bluetooth: Fallback from eSCO to SCO on unspecified error

Some Bluetooth chips (like the ones from Texas Instruments) don't do
proper eSCO negotiations inside the Link Manager. They just return an
error code and in case of the Kyocera ED-8800 headset it is just a
random error.

  < HCI Command: Setup Synchronous Connection 0x01|0x0028) plen 17
    handle 1 voice setting 0x0060
  > HCI Event: Command Status (0x0f) plen 4
    Setup Synchronous Connection (0x01|0x0028) status 0x00 ncmd 1
  > HCI Event: Synchronous Connect Complete (0x2c) plen 17
    status 0x1f handle 257 bdaddr 00:14:0A:xx:xx:xx type eSCO
    Error: Unspecified Error

In these cases it is up to the host stack to fallback to a SCO setup
and so retry with SCO parameters.

Based on a report by Nick Pelly <npelly@google.com>

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
15 years agoBluetooth: Fix removing of RFCOMM DLC timer with DEFER_SETUP
Johan Hedberg [Thu, 26 Mar 2009 14:41:56 +0000 (16:41 +0200)]
Bluetooth: Fix removing of RFCOMM DLC timer with DEFER_SETUP

There is a missing call to rfcomm_dlc_clear_timer in the case that
DEFER_SETUP is used and so the connection gets disconnected after the
timeout even if it was successfully accepted previously.

This patch adds a call to rfcomm_dlc_clear_timer to rfcomm_dlc_accept
which will get called when the user accepts the connection by calling
read() on the socket.

Signed-off-by: Johan Hedberg <johan.hedberg@nokia.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
15 years agovlan/macvlan: fix NULL pointer dereferences in ethtool handlers
Patrick McHardy [Fri, 17 Apr 2009 22:59:23 +0000 (15:59 -0700)]
vlan/macvlan: fix NULL pointer dereferences in ethtool handlers

Check whether the underlying device provides a set of ethtool ops before
checking for individual handlers to avoid NULL pointer dereferences.

Reported-by: Art van Breemen <ard@telegraafnet.nl>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wirel...
David S. Miller [Fri, 17 Apr 2009 22:54:40 +0000 (15:54 -0700)]
Merge branch 'master' of git://git./linux/kernel/git/linville/wireless-2.6

15 years agomyri10ge: fix compile error
Andrew Gallatin [Fri, 17 Apr 2009 22:45:15 +0000 (15:45 -0700)]
myri10ge: fix compile error

A compilation error snuck into
2d90b0aa3bc484189940444bcddcbe0ebbb53af5
due to an over-zealous indent script removing spaces around array
initialization ellipsis.  The attached patch fixes the myri10ge
compilation in net-next.

Signed-off-by: Andrew Gallatin <gallatin@myri.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/kaber/nf-2.6
David S. Miller [Fri, 17 Apr 2009 22:38:38 +0000 (15:38 -0700)]
Merge branch 'master' of git://git./linux/kernel/git/kaber/nf-2.6

15 years agoar9170usb: add ZyXEL NWD271N
Christian Lamparter [Fri, 17 Apr 2009 12:52:23 +0000 (14:52 +0200)]
ar9170usb: add ZyXEL NWD271N

Signed-off-by: Christian Lamparter <chunkeey@web.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agomac80211: validate TIM IE length
Johannes Berg [Thu, 16 Apr 2009 22:54:23 +0000 (00:54 +0200)]
mac80211: validate TIM IE length

The TIM IE must not be shorter than 4 bytes, so verify that
when parsing it.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agocfg80211: do not replace BSS structs
Johannes Berg [Thu, 16 Apr 2009 13:00:58 +0000 (15:00 +0200)]
cfg80211: do not replace BSS structs

Instead, allocate extra IE memory if necessary. Normally,
this isn't even necessary since there's enough space.

This is a better way of correcting the "held BSS can
disappear" issue, but also a lot more code. It is also
necessary for proper auth/assoc BSS handling in the
future.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agocfg80211: copy hold when replacing BSS
Johannes Berg [Thu, 16 Apr 2009 10:15:38 +0000 (12:15 +0200)]
cfg80211: copy hold when replacing BSS

When we receive a probe response frame we can replace the
BSS struct in our list -- but if that struct is held then
we need to hold the new one as well.

We really should fix this completely and not replace the
struct, but this is a bandaid for now.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agomac80211: avoid crashing when no scan sdata
Johannes Berg [Wed, 15 Apr 2009 19:33:15 +0000 (21:33 +0200)]
mac80211: avoid crashing when no scan sdata

Using the scan_sdata variable here is terribly wrong,
if there has never been a scan then we fail. However,
we need a bandaid...

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Cc: stable@kernel.org [2.6.29]
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agonetfilter: nfnetlink: return ENOMEM if we fail to create netlink socket
Pablo Neira Ayuso [Fri, 17 Apr 2009 15:48:44 +0000 (17:48 +0200)]
netfilter: nfnetlink: return ENOMEM if we fail to create netlink socket

With this patch, nfnetlink returns -ENOMEM instead of -EPERM if we
fail to create the nfnetlink netlink socket during the module
loading. This is exactly what rtnetlink does in this case.

Ideally, it would be better if we propagate the error that has
happened in netlink_kernel_create(), however, this function still
does not implement this yet.

Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Patrick McHardy <kaber@trash.net>
15 years agonetfilter: ctnetlink: report error if event message allocation fails
Pablo Neira Ayuso [Fri, 17 Apr 2009 15:47:31 +0000 (17:47 +0200)]
netfilter: ctnetlink: report error if event message allocation fails

This patch fixes an inconsistency that results in no error reports
to user-space listeners if we fail to allocate the event message.

Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Patrick McHardy <kaber@trash.net>
15 years ago[PATCH] net: remove superfluous call to synchronize_net()
Eric Dumazet [Fri, 17 Apr 2009 11:52:48 +0000 (04:52 -0700)]
[PATCH] net: remove superfluous call to synchronize_net()

inet_register_protosw() function is responsible for adding a new
inet protocol into a global table (inetsw[]) that is used with RCU rules.

As soon as the store of the pointer is done, other cpus might see
this new protocol in inetsw[], so we have to make sure new protocol
is ready for use. All pending memory updates should thus be committed
to memory before setting the pointer.
This is correctly done using rcu_assign_pointer()

synchronize_net() is typically used at unregister time, after
unsetting the pointer, to make sure no other cpu is still using
the object we want to dismantle. Using it at register time
is only adding an artificial delay that could hide a real bug,
and this bug could popup if/when synchronize_rcu() can proceed
faster than now.

This saves about 13 ms on boot time on a HZ=1000 8 cpus machine  ;)
(4 calls to inet_register_protosw(), and about 3200 us per call)

Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agogro: Fix use after free in tcp_gro_receive
Herbert Xu [Fri, 17 Apr 2009 09:34:38 +0000 (02:34 -0700)]
gro: Fix use after free in tcp_gro_receive

After calling skb_gro_receive skb->len can no longer be relied
on since if the skb was merged using frags, then its pages will
have been removed and the length reduced.

This caused tcp_gro_receive to prematurely end merging which
resulted in suboptimal performance with ixgbe.

The fix is to store skb->len on the stack.

Reported-by: Mark Wagner <mwagner@redhat.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agopegasus: Handle disconnect error code correctly.
Oliver Neukum [Fri, 17 Apr 2009 08:40:19 +0000 (01:40 -0700)]
pegasus: Handle disconnect error code correctly.

EPERM means that disconnect() is runnung. It should be treated like
ENODEV

Signed-off-by: Oliver Neukum <oliver@neukum.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agocan: Network Drop Monitor: Make use of consume_skb() in af_can.c
Oliver Hartkopp [Fri, 17 Apr 2009 08:38:46 +0000 (01:38 -0700)]
can: Network Drop Monitor: Make use of consume_skb() in af_can.c

Since commit ead2ceb0ec9f85cff19c43b5cdb2f8a054484431 ("Network Drop
Monitor: Adding kfree_skb_clean for non-drops and modifying
end-of-line points for skbs") so called end-of-line points for skb's
should use consume_skb() to free the socket buffer.

In opposite to consume_skb() the function kfree_skb() is intended to
be used for unexpected skb drops e.g. in error conditions that now can
trigger the network drop monitor if enabled.

This patch moves the skb end-of-line point in af_can.c to use
consume_skb().

Signed-off-by: Oliver Hartkopp <oliver@hartkopp.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agomacb: process the RX ring regardless of interrupt status
Erik Waling [Wed, 15 Apr 2009 23:32:11 +0000 (23:32 +0000)]
macb: process the RX ring regardless of interrupt status

Suppose that we receive lots of frames, start processing them, but
exhaust our budget so that we return before we had a chance to look
at all of them.

Then, when the network layer calls us again, we will only continue
processing the buffers if the REC bit was set in the mean time, which it
might not be if there was a brief pause in the flow of packets. If this
happens, we'll simply display a warning and call netif_rx_complete()
with potentially lots of unprocessed packets in the RX ring...

Fix this by scanning the ring no matter what flags are set in the
interrupt status register.

Signed-off-by: Erik Waling <erik.waling@konftel.com>
Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agomacb: Handle Retry Limit Exceeded errors
Erik Waling [Wed, 15 Apr 2009 23:32:10 +0000 (23:32 +0000)]
macb: Handle Retry Limit Exceeded errors

When transfering large amounts of data we sometimes experienced that the
Retry Limit Exceeded (RLE) bit got set in TSR during transmission
attempts. When this happened the driver would stall in a state that
prevented any more data from being sent.

Signed-off-by: Erik Waling <erik.waling@konftel.com>
Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoixgbe: Fix the DCB PFC thresholds for 82599
PJ Waskiewicz [Thu, 16 Apr 2009 15:00:41 +0000 (15:00 +0000)]
ixgbe: Fix the DCB PFC thresholds for 82599

The thresholds for the DCB priority flow control are incorrect for 82599.
This fixes the thresholds to be correct.

Signed-off-by: Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoixgbe: Fix DCB traffic class mapping for 82599
PJ Waskiewicz [Thu, 16 Apr 2009 15:00:20 +0000 (15:00 +0000)]
ixgbe: Fix DCB traffic class mapping for 82599

The traffic classes in hardware are not symmetrical for Rx and Tx.  Rx
is every 16 descriptor queues, Tx is not.  It runs 32-32-16-16-8-8-8 when
running with 8 traffic classes, and runs 64-32-16 when running with 4
traffic classes.  This patch fixes the mapping.

Signed-off-by: Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Cc: stable@kernel.org
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agowan/pc300_drv: convert to net_device_ops
Alexander Beregalov [Thu, 16 Apr 2009 15:48:17 +0000 (15:48 +0000)]
wan/pc300_drv: convert to net_device_ops

On Fri, Apr 17, 2009 at 05:23:02AM +0400, Alexander Beregalov wrote:
>
> Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
> ---
>  drivers/net/wan/pc300_drv.c |   22 ++++++++++++++--------
>  1 files changed, 14 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/net/wan/pc300_drv.c b/drivers/net/wan/pc300_drv.c
> index c23fde0..df10a4c 100644
> --- a/drivers/net/wan/pc300_drv.c
> +++ b/drivers/net/wan/pc300_drv.c
> @@ -225,6 +225,7 @@ static char rcsid[] =
>  #include <linux/skbuff.h>
>  #include <linux/if_arp.h>
>  #include <linux/netdevice.h>
> +#include <linux/etherdevice.h>
>  #include <linux/spinlock.h>
>  #include <linux/if.h>
>  #include <net/arp.h>
> @@ -3246,6 +3247,18 @@ static inline void show_version(void)
>   rcsvers, rcsdate, __DATE__, __TIME__);
>  } /* show_version */
>
> +static const struct net_device_ops cpc_netdev_ops = {
> + .ndo_init = NULL,
> + .ndo_open = cpc_open,
> + .ndo_stop = cpc_close,
> + .ndo_tx_timeout = cpc_tx_timeout,
> + .ndo_set_multicast_list = NULL,

In this case ndo_init and ndo_set_multicast_list
are not needed.

>From 1507a5a797a5f0005696a9bf10e390caca9c3800 Mon Sep 17 00:00:00 2001
From: Alexander Beregalov <a.beregalov@gmail.com>
Date: Fri, 17 Apr 2009 05:45:48 +0400
Subject: [PATCH] wan/pc300_drv: convert to net_device_ops

Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoirda: include etherdevice.h for eth_*() functions
Alexander Beregalov [Thu, 16 Apr 2009 15:23:03 +0000 (15:23 +0000)]
irda: include etherdevice.h for eth_*() functions

Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoe1000: fix transmit routine exit bug
Jesse Brandeburg [Thu, 16 Apr 2009 16:59:47 +0000 (16:59 +0000)]
e1000: fix transmit routine exit bug

If the e1000 transmit cleanup inner loop exited early, then
cleaned might not be true.  This could cause tx hangs or other
badness.  Use count to track the total number of descriptors
cleaned instead of basing a tx queue restart off of a temporary
working state variable.

This code now makes the flow the same for e1000/e1000e/igb/ixgbe

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoe1000e: fix bug in restart queue logic
Jesse Brandeburg [Thu, 16 Apr 2009 16:59:28 +0000 (16:59 +0000)]
e1000e: fix bug in restart queue logic

If the e1000e transmit cleanup inner loop exited early, then
cleaned might not be true.  This could cause tx hangs or other
badness.  Use count to track the total number of descriptors
cleaned instead of basing a tx queue restart off of a temporary
working state variable.

This code now makes the flow the same for e1000/e1000e/igb/ixgbe

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agomyri10ge: fix tx ring size in ethtool -g
Brice Goglin [Fri, 17 Apr 2009 00:56:57 +0000 (17:56 -0700)]
myri10ge: fix tx ring size in ethtool -g

Fix tx ring size reported via ethtool -g.

Signed-off-by: Brice Goglin <brice@myri.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agomyri10ge: add MODULE_DEVICE_TABLE
Brice Goglin [Thu, 16 Apr 2009 02:29:22 +0000 (02:29 +0000)]
myri10ge: add MODULE_DEVICE_TABLE

Add MODULE_DEVICE_TABLE so that modinfo reports pci device id aliases.

Signed-off-by: Brice Goglin <brice@myri.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agomyri10ge: allow per-board firmware overriding
Brice Goglin [Thu, 16 Apr 2009 02:24:59 +0000 (02:24 +0000)]
myri10ge: allow per-board firmware overriding

Add myri10ge_fw_names to override the default firmware
on a per-board basis.

Signed-off-by: Brice Goglin <brice@myri.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agomyri10ge: force stats update in ethtool gstats
Brice Goglin [Thu, 16 Apr 2009 02:23:56 +0000 (02:23 +0000)]
myri10ge: force stats update in ethtool gstats

Force a statistics update when our ethtool gstats routine
is called.  Otherwise, ethtool will continue to read stale
stats until something forces an update by reading /proc/net/dev

This requires putting a lock around the stats update to guard
against 2 threads (one via ethtool, and one via procfs)
updating the stats at once.

Signed-off-by: Brice Goglin <brice@myri.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoMAINTAINERS - Update NETWORKING W: url and standardize T: git url
Joe Perches [Thu, 16 Apr 2009 09:38:46 +0000 (09:38 +0000)]
MAINTAINERS - Update NETWORKING W: url and standardize T: git url

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoMAINTAINERS - Add section NETWORKING DRIVERS
Joe Perches [Thu, 16 Apr 2009 09:38:45 +0000 (09:38 +0000)]
MAINTAINERS - Add section NETWORKING DRIVERS

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agoMerge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
David S. Miller [Fri, 17 Apr 2009 00:35:26 +0000 (17:35 -0700)]
Merge branch 'master' of /linux/kernel/git/davem/net-2.6

15 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/kaber/nf-2.6
David S. Miller [Thu, 16 Apr 2009 23:32:29 +0000 (16:32 -0700)]
Merge branch 'master' of git://git./linux/kernel/git/kaber/nf-2.6

15 years agoixgbe: Be explict with what we are !'ing in ixgbe_sfp_config_module_task()
Tony Breeds [Thu, 16 Apr 2009 23:27:23 +0000 (16:27 -0700)]
ixgbe: Be explict with what we are !'ing in ixgbe_sfp_config_module_task()

GCC warns:
drivers/net/ixgbe/ixgbe_main.c: In function 'ixgbe_sfp_config_module_task':
drivers/net/ixgbe/ixgbe_main.c:3920: warning: suggest parantheses around
operand of '!' or change '&' to '&&' or '!' to '~'

Which I think is right.  Bracket to remove ambiguity.

Signed-off-by: Tony Breeds <tony@bakeyournoodle.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
15 years agonetfilter: nf_nat: add support for persistent mappings
Patrick McHardy [Thu, 16 Apr 2009 16:33:01 +0000 (18:33 +0200)]
netfilter: nf_nat: add support for persistent mappings

The removal of the SAME target accidentally removed one feature that is
not available from the normal NAT targets so far, having multi-range
mappings that use the same mapping for each connection from a single
client. The current behaviour is to choose the address from the range
based on source and destination IP, which breaks when communicating
with sites having multiple addresses that require all connections to
originate from the same IP address.

Introduce a IP_NAT_RANGE_PERSISTENT option that controls whether the
destination address is taken into account for selecting addresses.

http://bugzilla.kernel.org/show_bug.cgi?id=12954

Signed-off-by: Patrick McHardy <kaber@trash.net>
15 years agoar9170: fix struct layout on arm
Johannes Berg [Fri, 10 Apr 2009 08:02:45 +0000 (10:02 +0200)]
ar9170: fix struct layout on arm

arm will pad even between u8's, so mark the structs/unions
packed. Fixes a build bug on arm due to BUILD_BUG_ON tests
in the code.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Reported-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agoorinoco: correct timeout logic in __orinoco_hw_set_tkip_key()
Pavel Roskin [Fri, 10 Apr 2009 01:41:05 +0000 (21:41 -0400)]
orinoco: correct timeout logic in __orinoco_hw_set_tkip_key()

If the value read from HERMES_RID_TXQUEUEEMPTY becomes 0 after exactly
100 readings, we wrongly consider it a timeout.  Rewrite the clever
while loop as a for loop that does the right thing and looks simpler.

Reported by Juha Leppanen <juha_motorsportcom@luukku.com>

Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agomac80211: Fragmentation threshold (typo)
Gerrit Renker [Tue, 14 Apr 2009 04:32:56 +0000 (06:32 +0200)]
mac80211: Fragmentation threshold (typo)

mac80211: Fragmentation threshold (typo)

ieee80211_ioctl_siwfrag() sets the fragmentation_threshold to 2352
when frame fragmentation is to be disabled, yet the corresponding
'get' function tests for 2353 bytes instead.

This causes user-space tools to display a fragmentation threshold
of 2352 bytes even if fragmentation has been disabled.

Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agort2x00: prevent double kfree when failing to register hardware
Herton Ronaldo Krzesinski [Fri, 10 Apr 2009 21:05:14 +0000 (18:05 -0300)]
rt2x00: prevent double kfree when failing to register hardware

In a scenario where there isn't any firmware available, we will have a
double kfree of rt2x00dev->spec.channels_info when ieee80211_register_hw
returns an error status inside rt2x00lib_probe_hw.

The problem is that if ieee80211_register_hw fails, we call
rt2x00lib_remove_hw twice:
* first inside rt2x00lib_probe_hw upon failure of ieee80211_register_hw
* error status is returned to rt2x00lib_probe_dev, which then sees it and
  calls in this case rt2x00lib_remove_dev that will again run
  rt2x00lib_remove_hw

Prevent this avoiding calling rt2x00lib_remove_hw inside
rt2x00lib_probe_hw

Problem was detected with CONFIG_DEBUG_PAGEALLOC=y, CONFIG_SLUB_DEBUG=y,
CONFIG_SLUB_DEBUG_ON=y, that dumps this with no firmware available:

rt61pci 0000:00:07.0: PCI INT A -> GSI 19 (level, low) -> IRQ 19
wmaster0 (rt61pci): not using net_device_ops yet
phy0: Selected rate control algorithm 'pid'
phy0: Failed to initialize wep: -2
phy0 -> rt2x00lib_probe_dev: Error - Failed to initialize hw.
=============================================================================
BUG kmalloc-128: Object already free
-----------------------------------------------------------------------------

INFO: Allocated in rt61pci_probe_hw+0x3e5/0x6e0 [rt61pci] age=340 cpu=0 pid=21
INFO: Freed in rt2x00lib_remove_hw+0x59/0x70 [rt2x00lib] age=0 cpu=0 pid=21
INFO: Slab 0xc13ac3e0 objects=23 used=10 fp=0xdd59f6e0 flags=0x400000c3
INFO: Object 0xdd59f6e0 @offset=1760 fp=0xdd59f790

Bytes b4 0xdd59f6d0:  15 00 00 00 b2 8a fb ff 5a 5a 5a 5a 5a 5a 5a 5a ....².ûÿZZZZZZZZ
  Object 0xdd59f6e0:  6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b kkkkkkkkkkkkkkkk
  Object 0xdd59f6f0:  6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b kkkkkkkkkkkkkkkk
  Object 0xdd59f700:  6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b kkkkkkkkkkkkkkkk
  Object 0xdd59f710:  6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b kkkkkkkkkkkkkkkk
  Object 0xdd59f720:  6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b kkkkkkkkkkkkkkkk
  Object 0xdd59f730:  6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b kkkkkkkkkkkkkkkk
  Object 0xdd59f740:  6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b kkkkkkkkkkkkkkkk
  Object 0xdd59f750:  6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b a5 kkkkkkkkkkkkkkk¥
 Redzone 0xdd59f760:  bb bb bb bb                                     »»»»
 Padding 0xdd59f788:  5a 5a 5a 5a 5a 5a 5a 5a                         ZZZZZZZZ
Pid: 21, comm: stage1 Not tainted 2.6.29.1-desktop-1.1mnb #1
Call Trace:
 [<c01abbb3>] print_trailer+0xd3/0x120
 [<c01abd37>] object_err+0x37/0x50
 [<c01acf57>] __slab_free+0xe7/0x2f0
 [<c01ad1de>] kfree+0x7e/0xd0
 [<e0e4a239>] ? rt2x00lib_remove_hw+0x59/0x70 [rt2x00lib]
 [<e0e4a239>] ? rt2x00lib_remove_hw+0x59/0x70 [rt2x00lib]
 [<e0e4a239>] rt2x00lib_remove_hw+0x59/0x70 [rt2x00lib]
 [<e0e4acc7>] rt2x00lib_remove_dev+0x37/0x50 [rt2x00lib]
 [<e0e4b087>] rt2x00lib_probe_dev+0x1a7/0x3b0 [rt2x00lib]
 [<e0eb288f>] rt2x00pci_probe+0xdf/0x1ee [rt2x00pci]
 [<c026b9ee>] local_pci_probe+0xe/0x10
 [<c026c750>] pci_device_probe+0x60/0x80
 [<c02d5c2a>] driver_probe_device+0x9a/0x2e0
 [<c02d5ef9>] __driver_attach+0x89/0x90
 [<c02d541b>] bus_for_each_dev+0x4b/0x70
 [<c026c690>] ? pci_device_remove+0x0/0x40
 [<c02d59d9>] driver_attach+0x19/0x20
 [<c02d5e70>] ? __driver_attach+0x0/0x90
 [<c02d4cef>] bus_add_driver+0x1cf/0x2a0
 [<c026c690>] ? pci_device_remove+0x0/0x40
 [<c02d60c9>] driver_register+0x69/0x140
 [<c026c9b0>] __pci_register_driver+0x40/0x80
 [<e0ecc000>] ? rt61pci_init+0x0/0x19 [rt61pci]
 [<e0ecc017>] rt61pci_init+0x17/0x19 [rt61pci]
 [<c0101116>] do_one_initcall+0x26/0x1c0
 [<c01ab90c>] ? slab_pad_check+0x3c/0x120
 [<c01ab90c>] ? slab_pad_check+0x3c/0x120
 [<c01ac8da>] ? check_object+0xda/0x210
 [<c01b0026>] ? percpu_free+0x46/0x50
 [<c01ad09e>] ? __slab_free+0x22e/0x2f0
 [<c01b0026>] ? percpu_free+0x46/0x50
 [<c01b0026>] ? percpu_free+0x46/0x50
 [<c01b0026>] ? percpu_free+0x46/0x50
 [<c01687ec>] ? stop_machine_destroy+0x3c/0x40
 [<c015e515>] ? load_module+0xa5/0x1c50
 [<e0ec5000>] ? rt61pci_eepromregister_read+0x0/0x40 [rt61pci]
 [<e0eb2000>] ? rt2x00pci_write_tx_data+0x0/0x90 [rt2x00pci]
 [<c03ac2fb>] ? mutex_lock+0xb/0x20
 [<c03ac2fb>] ? mutex_lock+0xb/0x20
 [<c017ad16>] ? tracepoint_update_probe_range+0x76/0xa0
 [<c017ad6f>] ? tracepoint_module_notify+0x2f/0x40
 [<c03b02ed>] ? notifier_call_chain+0x2d/0x70
 [<c014f0ed>] ? __blocking_notifier_call_chain+0x4d/0x60
 [<c014f11a>] ? blocking_notifier_call_chain+0x1a/0x20
 [<c0160156>] sys_init_module+0x96/0x1d0
 [<c019dad6>] ? sys_munmap+0x46/0x60
 [<c0105546>] syscall_call+0x7/0xb
FIX kmalloc-128: Object at 0xdd59f6e0 not freed
rt61pci 0000:00:07.0: PCI INT A disabled
rt61pci: probe of 0000:00:07.0 failed with error -2

Signed-off-by: Herton Ronaldo Krzesinski <herton@mandriva.com.br>
Acked-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agofix iwl3945 registration regression
Johannes Berg [Thu, 9 Apr 2009 07:56:02 +0000 (09:56 +0200)]
fix iwl3945 registration regression

I forgot that iwl3945 registration is separate from iwlagn.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agomac80211: quiet beacon loss messages
Michael Buesch [Wed, 15 Apr 2009 18:41:22 +0000 (14:41 -0400)]
mac80211: quiet beacon loss messages

On Sunday 05 April 2009 11:29:38 Michael Buesch wrote:
> On Sunday 05 April 2009 11:23:59 Jaswinder Singh Rajput wrote:
> > With latest linus tree I am getting, .config file attached:
> >
> > [   22.895051] r8169: eth0: link down
> > [   22.897564] ADDRCONF(NETDEV_UP): eth0: link is not ready
> > [   22.928047] ADDRCONF(NETDEV_UP): wlan0: link is not ready
> > [   22.982292] libvirtd used greatest stack depth: 4200 bytes left
> > [   63.709879] wlan0: authenticate with AP 00:11:95:9e:df:f6
> > [   63.712096] wlan0: authenticated
> > [   63.712127] wlan0: associate with AP 00:11:95:9e:df:f6
> > [   63.726831] wlan0: RX AssocResp from 00:11:95:9e:df:f6 (capab=0x471 status=0 aid=1)
> > [   63.726855] wlan0: associated
> > [   63.730093] ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready
> > [   74.296087] wlan0: no IPv6 routers present
> > [   79.349044] wlan0: beacon loss from AP 00:11:95:9e:df:f6 - sending probe request
> > [  119.358200] wlan0: beacon loss from AP 00:11:95:9e:df:f6 - sending probe request
> > [  179.354292] wlan0: beacon loss from AP 00:11:95:9e:df:f6 - sending probe request
> > [  259.366044] wlan0: beacon loss from AP 00:11:95:9e:df:f6 - sending probe request
> > [  359.348292] wlan0: beacon loss from AP 00:11:95:9e:df:f6 - sending probe request
> > [  361.953459] packagekitd used greatest stack depth: 4160 bytes left
> > [  478.824258] wlan0: beacon loss from AP 00:11:95:9e:df:f6 - sending probe request
> > [  598.813343] wlan0: beacon loss from AP 00:11:95:9e:df:f6 - sending probe request
> > [  718.817292] wlan0: beacon loss from AP 00:11:95:9e:df:f6 - sending probe request
> > [  838.824567] wlan0: beacon loss from AP 00:11:95:9e:df:f6 - sending probe request
> > [  958.815402] wlan0: beacon loss from AP 00:11:95:9e:df:f6 - sending probe request
> > [ 1078.848434] wlan0: beacon loss from AP 00:11:95:9e:df:f6 - sending probe request
> > [ 1198.822913] wlan0: beacon loss from AP 00:11:95:9e:df:f6 - sending probe request
> > [ 1318.824931] wlan0: beacon loss from AP 00:11:95:9e:df:f6 - sending probe request
> > [ 1438.814157] wlan0: beacon loss from AP 00:11:95:9e:df:f6 - sending probe request
> > [ 1558.827336] wlan0: beacon loss from AP 00:11:95:9e:df:f6 - sending probe request
> > [ 1678.823011] wlan0: beacon loss from AP 00:11:95:9e:df:f6 - sending probe request
> > [ 1798.830589] wlan0: beacon loss from AP 00:11:95:9e:df:f6 - sending probe request
> > [ 1918.828044] wlan0: beacon loss from AP 00:11:95:9e:df:f6 - sending probe request
> > [ 2038.827224] wlan0: beacon loss from AP 00:11:95:9e:df:f6 - sending probe request
> > [ 2116.517152] wlan0: beacon loss from AP 00:11:95:9e:df:f6 - sending probe request
> > [ 2158.840243] wlan0: beacon loss from AP 00:11:95:9e:df:f6 - sending probe request
> > [ 2278.827427] wlan0: beacon loss from AP 00:11:95:9e:df:f6 - sending probe request
>
>
> I think this message should only show if CONFIG_MAC80211_VERBOSE_DEBUG is set.
> It's kind of expected that we lose a beacon once in a while, so we shouldn't print
> verbose messages to the kernel log (even if they are KERN_DEBUG).
>
> And besides that, I think one can easily remotely trigger this message and flood the logs.
> So it should probably _also_ be ratelimited.

Something like this:

Signed-off-by: Michael Buesch <mb@bu3sch.de>
15 years agob43: Do radio lock assertion in software
Michael Buesch [Tue, 31 Mar 2009 10:27:32 +0000 (12:27 +0200)]
b43: Do radio lock assertion in software

The assertion of the lock-bit in the hardware register is unreliable,
because there are devices with quirks that will randomly set the bit.

Do the assertion in software, only.

Signed-off-by: Michael Buesch <mb@bu3sch.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agort2x00: Add rt73usb USB IDs
Ivo van Doorn [Sat, 28 Mar 2009 19:52:14 +0000 (20:52 +0100)]
rt2x00: Add rt73usb USB IDs

Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agop54usb: add Telsey 802.11g USB2.0 Adapter
Christian Lamparter [Sun, 29 Mar 2009 20:50:28 +0000 (22:50 +0200)]
p54usb: add Telsey 802.11g USB2.0 Adapter

Signed-off-by: Christian Lamparter <chunkeey@web.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agoar9170: add Cace Airpcap NX usb_id
Christian Lamparter [Tue, 24 Mar 2009 20:58:08 +0000 (21:58 +0100)]
ar9170: add Cace Airpcap NX usb_id

This patch adds a new device to ar9170usb.

Reported-by: Mike Kershaw/Dragorn <dragorn@kismetwireless.net>
Signed-off-by: Christian Lamparter <chunkeey@web.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agoairo: queue SIOCSIWAUTH-requested auth mode change for next commit
Dan Williams [Wed, 8 Apr 2009 14:15:17 +0000 (10:15 -0400)]
airo: queue SIOCSIWAUTH-requested auth mode change for next commit

Code was clearly wrong, plus callers expect the mode change to happen as
soon as possible, not dropped on the floor until the next time some
other config value changes and a commit happens.

Signed-off-by: Dan Williams <dcbw@redhat.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
15 years agoat76c50x-usb: Add device ID for OQO model 01+
Jamie Lentin [Tue, 7 Apr 2009 21:59:56 +0000 (22:59 +0100)]
at76c50x-usb: Add device ID for OQO model 01+

Add USB device ID for OQO 01+'s internal wireless LAN

An OQO employee mentions the chip's true identity here:-
    ftp://ftp.oqo.com/unsupported/linux/OQOLinux.html

Signed-off-by: Jamie Lentin <jm@lentin.co.uk>
Acked-by: Kalle Valo <kalle.valo@iki.fi>
Signed-off-by: John W. Linville <linville@tuxdriver.com>