The semantic patch I used was this:

@@
expression lock;
identifier flags;
expression subclass;
@@

- unsigned int flags;
+ unsigned long flags;

...

<+...

(
 spin_lock_irqsave(lock, flags)
|
 _spin_lock_irqsave(lock)
|
 spin_unlock_irqrestore(lock, flags)
|
 _spin_unlock_irqrestore(lock, flags)
|
 read_lock_irqsave(lock, flags)
|
 _read_lock_irqsave(lock)
|
 read_unlock_irqrestore(lock, flags)
|
 _read_unlock_irqrestore(lock, flags)
|
 write_lock_irqsave(lock, flags)
|
 _write_lock_irqsave(lock)
|
 write_unlock_irqrestore(lock, flags)
|
 _write_unlock_irqrestore(lock, flags)
|
 spin_lock_irqsave_nested(lock, flags, subclass)
|
 _spin_lock_irqsave_nested(lock, subclass)
|
 spin_unlock_irqrestore(lock, flags)
|
 _spin_unlock_irqrestore(lock, flags)
|
 _raw_spin_lock_flags(lock, flags)
|
 __raw_spin_lock_flags(lock, flags)
)

...+>

This patch was generated using the Coccinelle framework.

Cc: Masakazu Mokuno <mokuno@sm.sony.co.jp>
Cc: Julia Lawall <julia@diku.dk>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: NVegard Nossum <vegard.nossum@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This patch fixes mac80211 to not use the skb->cb over the queue step
from virtual interfaces to the master. The patch also, for now,
disables aggregation because that would still require requeuing,
will fix that in a separate patch. There are two other places (software
requeue and powersaving stations) where requeue can happen, but that is
not currently used by any drivers/not possible to use respectively.
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Wireless statistics produced by the RTL8187B driver are not particularly
informative about the strength of the received signal. From the data sheet
provided by Realtek, I discovered that certain parts of the RX header
should have the information necessary to calculate signal quality and
strength. With testing, it became clear that most of these quantities were
very jittery - only the AGC correlated with the signals expected from nearby
AP's. As a result, the quality and strength are derived from the agc value.
The scaling has been determined so that the numbers are close to those
obtained by b43 under the same conditions. The results are qualitatively
correct.

Statistics derived for the RTL8187 have not been changed.

The RX header variables have been renamed to match the quantites described
in the Realtek data sheet.
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

"mac80211: fix TX sequence numbers" broke rtl8187. This
patch makes the same kind of fix that was done for rt2x00. Note that
this code will have to be reworked for proper sequence numbers on beacons.
In addition, the sequence number has been placed in the hardware state,
not the vif state.
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

drivers/net/wireless/mac80211_hwsim.c:503:20: warning: Using plain integer as NULL pointer
Signed-off-by: NHarvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

ia64:

drivers/net/wireless/iwlwifi/iwl-led.c: In function `iwl_get_blink_rate':
drivers/net/wireless/iwlwifi/iwl-led.c:271: warning: long long int format, s64 arg (arg 6)
drivers/net/wireless/iwlwifi/iwl-led.c:271: warning: long long int format, u64 arg (arg 7)

We do not know what type the architecture uses to impement u64 and s64,
hence we must cast the variables for printing.

Cc: Tomas Winkler <tomas.winkler@intel.com>
Cc: John W. Linville <linville@tuxdriver.com>
Cc: Zhu Yi <yi.zhu@intel.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

When the queues are being initialized the entry flags fields must be
reset to 0. When this does not happen some entries might still be
marked as "occupied" after an ifdown & ifup cycle which would trigger
errors when the entry is being accessed:

	phy0 -> rt2x00queue_write_tx_frame: Error - Arrived at non-free entry in the non-full queue 0.
	Please file bug report to http://rt2x00.serialmonkey.com.

This also fixes the mac80211 warning:

	------------[ cut here ]------------
	WARNING: at net/mac80211/tx.c:1238 ieee80211_master_start_xmit+0x30a/0x350 [mac80211]()

which was triggered by the queue error.
Signed-off-by: NIvo van Doorn <IvDoorn@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

rt2x00 will only perform configuration changes from
mac80211 when the configuration option has changed.
This means it keeps track of the current active configuration
and will check these values when the config() callback function
is used.

However this causes breakage when the interface has been
brought down and up again, since all stored active values
aren't reset while the registers might have.
This is for example the case with rt61pci antenna registers which
will jump to invalid values when the interface has been started.

To make sure a full configuration takes place after the start()
callback function, a new flag is added which will be checked
during config() and skips the "what's changed" phase.
Signed-off-by: NIvo van Doorn <IvDoorn@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

The mutex is released on a successful return, so it would seem that it
should be released on an error return as well.

The semantic patch finds this problem is as follows:
(http://www.emn.fr/x-info/coccinelle/)

// <smpl>
@@
expression l;
@@

mutex_lock(l);
... when != mutex_unlock(l)
    when any
    when strict
(
if (...) { ... when != mutex_unlock(l)
+   mutex_unlock(l);
    return ...;
}
|
mutex_unlock(l);
)
// </smpl>
Signed-off-by: NJulia Lawall <julia@diku.dk>
Signed-off-by: NMichael Buesch <mb@bu3sch.de>
Cc: stable <stable@kernel.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

The mutex is released on a successful return, so it would seem that it
should be released on an error return as well.

The semantic patch finds this problem is as follows:
(http://www.emn.fr/x-info/coccinelle/)

// <smpl>
@@
expression l;
@@

mutex_lock(l);
... when != mutex_unlock(l)
    when any
    when strict
(
if (...) { ... when != mutex_unlock(l)
+   mutex_unlock(l);
    return ...;
}
|
mutex_unlock(l);
)
// </smpl>
Signed-off-by: NJulia Lawall <julia@diku.dk>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

When sending the RTS/CTS frame fails, we should
free the skb buffer which was created.
Signed-off-by: NIvo van Doorn <IvDoorn@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

When IEEE80211_TX_CTL_ASSIGN_SEQ is not set,
the driver should disable hardware sequence counting
to make sure the mac80211 provided counter is used.
This fixes QOS sequence counting, since that is one
of the cases where mac80211 provides a seperate
sequence counter.

By moving the sequence counting code to rt2x00queue
we make sure that _all_ frames get the sequence counter,
including RTS/CTS and Beacon frames.
Signed-off-by: NIvo van Doorn <IvDoorn@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

When setting up a beacon template, the length of the beacon is
calculated with the assumption that the SKB already contains
the Tx descriptor. In the case of beacons it doesn't.

This patch undoes the damage by adding the Tx descriptor length
to the beacon length. This is safe, because the shortest possible
beacon is longer than the Tx header.
Signed-off-by: NIwo Mergler <Iwo@call-direct.com.au>
Signed-off-by: NIvo van Doorn <IvDoorn@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Switches rt73usb to use large vendor requests for firmware
and beacons. This also fixes the garbled beacon bug.
Signed-off-by: NIwo Mergler <Iwo@call-direct.com.au>
Signed-off-by: NIvo van Doorn <IvDoorn@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Adds an extra rt2x00 vendor request function to support register
transfers beyond the CSR_CACHE_SIZE / USB packet size limit. This
is useful for firmware uploads, beacon templates and keys, all
of which are to large to do with a single USB request.
Signed-off-by: NIwo Mergler <Iwo@call-direct.com.au>
Signed-off-by: NIvo van Doorn <IvDoorn@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Olivier reported a difference between the EIFS
values used in the legacy driver and the one in
the rt2x00 drivers.

In rt2x00 the value was
	( SIFS + (8 * (IEEE80211_HEADER + ACK_SIZE)) )
which comes down to 314us while the legacy driver uses the value 364us

This was caused because EIFS is: SIFS + DIFS + AckTime
This patch will fix this by adding the DIFS by the above value,
and creating a SHORT_EIFS define which uses the SHORT_DIFS.
Reported-by: NOlivier Cornu <o.cornu@gmail.com>
Signed-off-by: NIvo van Doorn <IvDoorn@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

rt2x00queue_free_skb() was declared twice.
Signed-off-by: NIvo van Doorn <IvDoorn@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

MSI is a nice thing, but we cannot enable it without changing the
interrupt handler.  If we do it, we break MSI capable hardware,
specifically AR5006 chipset.
Signed-off-by: NPavel Roskin <proski@gnu.org>
Acked-by: NNick Kossifidis <mickflemm@gmail.com>
Cc: stable <stable@kernel.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

ath5k_beacon_update takes sc->lock upon entry.  However, it is only
called from within ath5k_config_interface, which already holds the lock.
Remove the unnecessary locking from ath5k_beacon_update.
Signed-off-by: NBob Copeland <me@bobcopeland.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

CC [M]  drivers/net/wireless/iwlwifi/iwl-scan.o
drivers/net/wireless/iwlwifi/iwl-scan.c: In function 'iwl_rx_scan_complete_notif':
drivers/net/wireless/iwlwifi/iwl-scan.c:274: warning: unused variable 'scan_notif'
Signed-off-by: NDenis V. Lunev <den@openvz.org>
Acked-by: NTomas Winkler <tomasw@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

CC [M]  drivers/net/wireless/iwlwifi/iwl-rfkill.o
drivers/net/wireless/iwlwifi/iwl-led.c: In function 'iwl_led_brightness_set':
drivers/net/wireless/iwlwifi/iwl-led.c:198: error: 'led_type_str' undeclared (first use in this function)
drivers/net/wireless/iwlwifi/iwl-led.c:198: error: (Each undeclared identifier is reported only once
drivers/net/wireless/iwlwifi/iwl-led.c:198: error: for each function it appears in.)

The problem is that led_type_str is defined under CONFIG_IWLWIFI_DEBUG
while IWL_DEBUG is a static inline function in this case. Replace it
with macro.
Signed-off-by: NDenis V. Lunev <den@openvz.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Some persistent settings were using hex and others decimal.  In some cases,
values were set in hex but reported in decimal.  Confusing.
Signed-off-by: NBrian Cavagnolo <brian@cozybit.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

When building the wireless-next-2.6 tree with CONFIG_IWL3945 (for building
iwl-3945 driver) and where CONFIG_IWL3945_LEDS is not set,
we get this warning:
drivers/net/wireless/iwlwifi/iwl-3945.c: In function
'iwl3945_pass_packet_to_mac80211':
drivers/net/wireless/iwlwifi/iwl-3945.c:633: warning: unused variable 'hdr'

This patch adds #ifdef to iwl3945_pass_packet_to_mac80211() to avoid this
warning. (The variable 'hdr' is used only if CONFIG_IWL3945_LEDS is set)
Signed-off-by: NRami Rosen <ramirose@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Helge Deller reports that HP laptops (NC4010 and NC6000) use active-
high signals to turn on the LEDs.  Previous code used active-low for
all devices.
Signed-off-by: NBob Copeland <me@bobcopeland.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

- free and re-request irq since it might have changed during suspend
- disable and enable msi
- don't set D0 state of the device, it's already done by the PCI layer
- do restore_state before enable_device, it's safer
- check ath5k_init return value
Signed-off-by: NJiri Slaby <jirislaby@gmail.com>
Acked-by: NNick Kossifidis <mickflemm@gmail.com>
Cc: Luis R. Rodriguez <mcgrof@gmail.com>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Don't sync
- coherent mapping (descriptors)
- before unmap, it's useless
- (wrongly anyway -- for_cpu) beacon skb, it's just mapped,
  so by the device yet
Signed-off-by: NJiri Slaby <jirislaby@gmail.com>
Acked-by: NNick Kossifidis <mickflemm@gmail.com>
Cc: Luis R. Rodriguez <mcgrof@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Make sure that the irq is not in progress after stop. This means
two things:
- ensure the intr setting register is set by flushing posted values
- call synchronize_irq() after that

Also flush stop tx write, inform callers of the tx stop about still
pending transfers (unsuccessful stop) and finally don't wait another
3ms in ath5k_rx_stop, since ath5k_hw_stop_rx_dma ensures transfer to
be finished.

Make sure all writes will be ordered in respect to locks by mmiowb().
Signed-off-by: NJiri Slaby <jirislaby@gmail.com>
Acked-by: NNick Kossifidis <mickflemm@gmail.com>
Cc: Luis R. Rodriguez <mcgrof@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Don't forget to kill tasklets on stop to not panic if they
fire after freeing some structures.
Signed-off-by: NJiri Slaby <jirislaby@gmail.com>
Acked-by: NNick Kossifidis <mickflemm@gmail.com>
Cc: Luis R. Rodriguez <mcgrof@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

When signal is noisy, hardware can use all RX buffers and since the last
entry in the list is self-linked, it overwrites the entry until we link
new buffers.

Ensure that we don't free this last one until we are 100% sure that it
is not used by the hardware anymore to not cause memory curruption as
can be seen below.

This is done by checking next buffer in the list. Even after that we
know that the hardware refetched the new link and proceeded further
(the next buffer is ready) we can finally free the overwritten buffer.

We discard it since the status in its descriptor is overwritten (OR-ed
by new status) too.

=============================================================================
BUG kmalloc-4096: Poison overwritten
-----------------------------------------------------------------------------

INFO: 0xffff810067419060-0xffff810067419667. First byte 0x8 instead of 0x6b
INFO: Allocated in dev_alloc_skb+0x18/0x30 age=1118 cpu=1 pid=0
INFO: Freed in skb_release_data+0x85/0xd0 age=1105 cpu=1 pid=3718
INFO: Slab 0xffffe200019d0600 objects=7 used=0 fp=0xffff810067419048 flags=0x40000000000020c3
INFO: Object 0xffff810067419048 @offset=4168 fp=0xffff81006741c120

Bytes b4 0xffff810067419038:  4f 0b 02 00 01 00 00 00 5a 5a 5a 5a 5a 5a 5a 5a O.......ZZZZZZZZ
  Object 0xffff810067419048:  6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b 6b kkkkkkkkkkkkkkkk
  Object 0xffff810067419058:  6b 6b 6b 6b 6b 6b 6b 6b 08 42 30 00 00 0b 6b 80 kkkkkkkk.B0...k.
  Object 0xffff810067419068:  f0 5d 00 4f 62 08 a3 64 00 0c 42 16 52 e4 f0 5a 360].Ob.243d..B.R344360Z
  Object 0xffff810067419078:  68 81 00 00 7b a5 b4 be 7d 3b 8f 53 cd d5 de 12 h...{245264276};.S315325336.
  Object 0xffff810067419088:  96 10 0b 89 48 54 23 41 0f 4e 2d b9 37 c3 cb 29 ....HT#A.N-2717303313)
  Object 0xffff810067419098:  d1 e0 de 14 8a 57 2a cc 3b 44 0d 78 7a 19 12 15 321340336..W*314;D.xz...
  Object 0xffff8100674190a8:  a9 ec d4 35 a8 10 ec 8c 40 a7 06 0a 51 a7 48 bb 2513543245250.354.@247..Q247H273
  Object 0xffff8100674190b8:  3e cf a1 c7 38 60 63 3f 51 15 c7 20 eb ba 65 30 >ϡ3078`c?Q.307.353272e0
 Redzone 0xffff81006741a048:  bb bb bb bb bb bb bb bb                         273273273273273273273273
 Padding 0xffff81006741a088:  5a 5a 5a 5a 5a 5a 5a 5a                         ZZZZZZZZ
Pid: 3297, comm: ath5k_pci Not tainted 2.6.26-rc8-mm1_64 #427

Call Trace:
 [<ffffffff802a7306>] print_trailer+0xf6/0x150
 [<ffffffff802a7485>] check_bytes_and_report+0x125/0x180
 [<ffffffff802a75dc>] check_object+0xac/0x260
 [<ffffffff802a9308>] __slab_alloc+0x368/0x6d0
 [<ffffffff80544f82>] ? wireless_send_event+0x142/0x310
 [<ffffffff804b1bd4>] ? __alloc_skb+0x44/0x150
 [<ffffffff80544f82>] ? wireless_send_event+0x142/0x310
 [<ffffffff802aa853>] __kmalloc_track_caller+0xc3/0xf0
 [<ffffffff804b1bfe>] __alloc_skb+0x6e/0x150
[... stack snipped]

FIX kmalloc-4096: Restoring 0xffff810067419060-0xffff810067419667=0x6b

FIX kmalloc-4096: Marking all objects used
Signed-off-by: NJiri Slaby <jirislaby@gmail.com>
Acked-by: NNick Kossifidis <mickflemm@gmail.com>
Cc: Luis R. Rodriguez <mcgrof@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

I forgot this in the previous patch that made it unused.
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

make it atomic_long_t; while we are at it, get rid of useless checks in affs,
hfs and hpfs - ->open() always has it equal to 1, ->release() - to 0.
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

If we can be sure that elevating the page_count on a pagecache page will
pin it, we can speculatively run this operation, and subsequently check to
see if we hit the right page rather than relying on holding a lock or
otherwise pinning a reference to the page.

This can be done if get_page/put_page behaves consistently throughout the
whole tree (ie.  if we "get" the page after it has been used for something
else, we must be able to free it with a put_page).

Actually, there is a period where the count behaves differently: when the
page is free or if it is a constituent page of a compound page.  We need
an atomic_inc_not_zero operation to ensure we don't try to grab the page
in either case.

This patch introduces the core locking protocol to the pagecache (ie.
adds page_cache_get_speculative, and tweaks some update-side code to make
it work).

Thanks to Hugh for pointing out an improvement to the algorithm setting
page_count to zero when we have control of all references, in order to
hold off speculative getters.

[kamezawa.hiroyu@jp.fujitsu.com: fix migration_entry_wait()]
[hugh@veritas.com: fix add_to_page_cache]
[akpm@linux-foundation.org: repair a comment]
Signed-off-by: NNick Piggin <npiggin@suse.de>
Cc: Jeff Garzik <jeff@garzik.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Hugh Dickins <hugh@veritas.com>
Cc: "Paul E. McKenney" <paulmck@us.ibm.com>
Reviewed-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: NDaisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Signed-off-by: NKAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: NKOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: NHugh Dickins <hugh@veritas.com>
Acked-by: NNick Piggin <npiggin@suse.de>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Add per-device dma_mapping_ops support for CONFIG_X86_64 as POWER
architecture does:

This enables us to cleanly fix the Calgary IOMMU issue that some devices
are not behind the IOMMU (http://lkml.org/lkml/2008/5/8/423).

I think that per-device dma_mapping_ops support would be also helpful for
KVM people to support PCI passthrough but Andi thinks that this makes it
difficult to support the PCI passthrough (see the above thread).  So I
CC'ed this to KVM camp.  Comments are appreciated.

A pointer to dma_mapping_ops to struct dev_archdata is added.  If the
pointer is non NULL, DMA operations in asm/dma-mapping.h use it.  If it's
NULL, the system-wide dma_ops pointer is used as before.

If it's useful for KVM people, I plan to implement a mechanism to register
a hook called when a new pci (or dma capable) device is created (it works
with hot plugging).  It enables IOMMUs to set up an appropriate
dma_mapping_ops per device.

The major obstacle is that dma_mapping_error doesn't take a pointer to the
device unlike other DMA operations.  So x86 can't have dma_mapping_ops per
device.  Note all the POWER IOMMUs use the same dma_mapping_error function
so this is not a problem for POWER but x86 IOMMUs use different
dma_mapping_error functions.

The first patch adds the device argument to dma_mapping_error.  The patch
is trivial but large since it touches lots of drivers and dma-mapping.h in
all the architecture.

This patch:

dma_mapping_error() doesn't take a pointer to the device unlike other DMA
operations.  So we can't have dma_mapping_ops per device.

Note that POWER already has dma_mapping_ops per device but all the POWER
IOMMUs use the same dma_mapping_error function.  x86 IOMMUs use device
argument.

[akpm@linux-foundation.org: fix sge]
[akpm@linux-foundation.org: fix svc_rdma]
[akpm@linux-foundation.org: build fix]
[akpm@linux-foundation.org: fix bnx2x]
[akpm@linux-foundation.org: fix s2io]
[akpm@linux-foundation.org: fix pasemi_mac]
[akpm@linux-foundation.org: fix sdhci]
[akpm@linux-foundation.org: build fix]
[akpm@linux-foundation.org: fix sparc]
[akpm@linux-foundation.org: fix ibmvscsi]
Signed-off-by: NFUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Cc: Muli Ben-Yehuda <muli@il.ibm.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Avi Kivity <avi@qumranet.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Removes legacy reinvent-the-wheel type thing. The generic
machinery integrates much better to automated debugging aids
such as kerneloops.org (and others), and is unambiguous due to
better naming. Non-intuively BUG_TRAP() is actually equal to
WARN_ON() rather than BUG_ON().
Signed-off-by: NIlpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Update existing Mellanox copyright lines to 2008, and add such lines
to files where they are missing.
Signed-off-by: NJack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: NRoland Dreier <rolandd@cisco.com>

Enable ibmveth for Cooperative Memory Overcommitment (CMO).  For this driver
it means calculating a desired amount of IO memory based on the current MTU
and updating this value with the bus when MTU changes occur.  Because DMA
mappings can fail, we have added a bounce buffer for temporary cases where
the driver can not map IO memory for the buffer pool.

The following changes are made to enable the driver for CMO:
 * DMA mapping errors will not result in error messages if entitlement has
   been exceeded and resources were not available.
 * DMA mapping errors are handled gracefully, ibmveth_replenish_buffer_pool()
   is corrected to check the return from dma_map_single and fail gracefully.
 * The driver will have a get_desired_dma function defined to function
   in a CMO environment.
 * When the MTU is changed, the driver will update the device IO entitlement
Signed-off-by: NRobert Jennings <rcj@linux.vnet.ibm.com>
Signed-off-by: NBrian King <brking@linux.vnet.ibm.com>
Signed-off-by: NSantiago Leon <santil@us.ibm.com>
Acked-by: NPaul Mackerras <paulus@samba.org>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Activates larger rx buffer pools when the MTU is changed to a larger
value.  This patch de-activates the large rx buffer pools when the MTU
changes to a smaller value.
Signed-off-by: NSantiago Leon <santil@us.ibm.com>
Signed-off-by: NRobert Jennings <rcj@linux.vnet.ibm.com>
Acked-by: NPaul Mackerras <paulus@samba.org>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

If we hack the virtio_net driver to always allocate full-sized (64k+)
skbuffs, the driver slows down (lguest numbers):

  Time to receive 1GB (small buffers): 10.85 seconds
  Time to receive 1GB (64k+ buffers): 24.75 seconds

Of course, large buffers use up more space in the ring, so we increase
that from 128 to 2048:

  Time to receive 1GB (64k+ buffers, 2k ring): 16.61 seconds

If we recycle pages rather than using alloc_page/free_page:

  Time to receive 1GB (64k+ buffers, 2k ring, recycle pages): 10.81 seconds

This demonstrates that with efficient allocation, we don't need to
have a separate "small buffer" queue.
Signed-off-by: NRusty Russell <rusty@rustcorp.com.au>

Finally this patch lets virtio_net receive GSO packets in addition
to sending them.  This can definitely be optimised for the non-GSO
case.  For comparison the Xen approach stores one page in each skb
and uses subsequent skb's pages to construct an SG skb instead of
preallocating the maximum amount of pages per skb.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au> (added feature bits)

This patch adds some basic ethtool operations to virtio_net so
I could test SG without GSO (which was really useful because TSO
turned out to be buggy :)

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au> (remove MTU setting)