The previous value of 672 for L2CAP_DEFAULT_MAX_PDU_SIZE is based on
the default L2CAP MTU.  That default MTU is calculated from the size
of two DH5 packets, minus ACL and L2CAP b-frame header overhead.

ERTM is used with newer basebands that typically support larger 3-DH5
packets, and i-frames and s-frames have more header overhead.  With
clean RF conditions, basebands will typically attempt to use 1021-byte
3-DH5 packets for maximum throughput.  Adjusting for 2 bytes of ACL
headers plus 10 bytes of worst-case L2CAP headers yields 1009 bytes
of payload.

This PDU size imposes less overhead for header bytes and gives the
baseband the option to choose 3-DH5 packets, but is small enough for
ERTM traffic to interleave well with other L2CAP or SCO data.
672-byte payloads do not allow the most efficient over-the-air
packet choice, and cannot achieve maximum throughput over BR/EDR.
Signed-off-by: Mat Martineau <mathewm@codeaurora.org>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

remote_tx_win is intended to be set on receipt of an L2CAP
configuration request.  The value is used to determine the size of the
transmit window on the remote side of an ERTM connection, so L2CAP
can stop sending frames when that remote window is full.

An incorrect remote_tx_win value will cause the stack to not fully
utilize the tx window (performance impact), or to overfill the remote
tx window (causing dropped frames or a disconnect).

This patch removes an extra setting of remote_tx_win when a
configuration response is received.  The transmit window has a
different meaning in a response - it is an informational value
less than or equal to the local tx_win.
Signed-off-by: Mat Martineau <mathewm@codeaurora.org>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

The L2CAP specification requires that the ERTM retransmit timeout be at
least 2 seconds for BR/EDR connections.
Signed-off-by: Mat Martineau <mathewm@codeaurora.org>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

Incoming configuration values must be converted to native CPU order
before use.  This fixes a bug where a little-endian MPS value is
compared to a native CPU value.  On big-endian processors, this
can cause ERTM and streaming mode segmentation to produce PDUs
that are larger than the remote stack is expecting, or that would
produce fragmented skbs that the current FCS code cannot handle.
Signed-off-by: Mat Martineau <mathewm@codeaurora.org>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>

This is based on work originally done by Patric McHardy.
Signed-off-by: Ben Greear <greearb@candelatech.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Driver should call pci_disable_device() if it returns from pci_probe()
with error.
Signed-off-by: Kulikov Vasiliy <segooon@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Driver should call pci_disable_device() if it returns from pci_probe()
with error.
Signed-off-by: Kulikov Vasiliy <segooon@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Driver should call pci_disable_device() if it returns from pci_probe()
with error.
Signed-off-by: Kulikov Vasiliy <segooon@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch makes rx_submit() return an error code, and makes some call sites
that care check the return value. This is important because it lets us properly
handle cases where the device isn't ready to handle URB submissions (e.g., when
it is autosuspended under some drivers); previously, we would attempt and fail
to submit URBs and reschedule ourselves to try and fail again. This patch is
against Linus's 2.6 repo commit 45d7f32c.
Signed-Off-By: Elizabeth Jones <ellyjones@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

pkt_sched: Add some basic qdisc class ops verification. Was: [PATCH] sfq: add dummy bind/unbind handles

Verify in register_qdisc() some basic qdisc class handlers are present.
Signed-off-by: Jarek Poplawski <jarkao2@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

pkt_sched: sch_sfq: Add dummy unbind_tcf and put handles. Was: [PATCH] sfq: add dummy bind/unbind handles

Add dummy .unbind_tcf and .put qdisc class ops for easier verification.
(All other schedulers have it like this.)
Signed-off-by: Jarek Poplawski <jarkao2@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

qlcnic_pci_info structs are 128 bytes so an array of 8 uses 1024 bytes.
That's a lot if you run with 4K stacks.  I allocated them with kcalloc()
instead.
Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

In the original code we allocated memory conditionally and freed it in
the error handling unconditionally.  It turns out that this function is
only called during initialization and "adapter->npars" and
"adapter->eswitch" are always NULL at the start of the function.  I
removed those checks.

Also since I was cleaning things, I changed the error handling for
qlcnic_get_pci_info() and pulled everything in an indent level.
Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Fix phy.c kernel-doc notation:

Warning(drivers/net/phy/phy.c:313): No description found for parameter 'ifr'
Warning(drivers/net/phy/phy.c:313): Excess function parameter 'mii_data' description in 'phy_mii_ioctl'
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add missing kernel-doc notation to struct sock:

Warning(include/net/sock.h:324): No description found for parameter 'sk_peer_pid'
Warning(include/net/sock.h:324): No description found for parameter 'sk_peer_cred'
Warning(include/net/sock.h:324): No description found for parameter 'sk_classid'
Warning(include/net/sock.h:324): Excess struct/union/enum/typedef member 'sk_peercred' description in 'sock'
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Fix etherdevice.h parameter name typo in kernel-doc:

Warning(include/linux/etherdevice.h:138): No description found for parameter 'hwaddr'
Warning(include/linux/etherdevice.h:138): Excess function parameter 'addr' description in 'dev_hw_addr_random'
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/wan/farsync.c: In function 'fst_intr_rx':
drivers/net/wan/farsync.c:1312: warning: cast to pointer from integer of different size
drivers/net/wan/farsync.c: In function 'do_bottom_half_tx':
drivers/net/wan/farsync.c:1407: warning: cast to pointer from integer of different size

The "skb" and "mem" arguments being passed here are DMA addresses
being programmed into the hardware registers, so pass them as the type
that they actually are.  And use the correct printf formatting in
debug logging statements for these things to match the type change.
Reported-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently the driver will try to protect all frames,
which leads to a lot of odd things like sending an
RTS with a zeroed RA before multicast frames, which
is clearly bogus.

In order to fix all of this, we need to take a step
back and see what we need to achieve:
 * we need RTS/CTS protection if requested by
   the AP for the BSS, mac80211 tells us this
 * in that case, CTS-to-self should only be
   enabled when mac80211 tells us
 * additionally, as a hardware workaround, on
   some devices we have to protect aggregated
   frames with RTS

To achieve the first two items, set up the RXON
accordingly and set the protection required flag
in the transmit command when mac80211 requests
protection for the frame.

To achieve the last item, set the rate-control
RTS-requested flag for all stations that we have
aggregation sessions with, and set the protection
required flag when sending aggregated frames (on
those devices where this is required).

Since otherwise bugs can occur, do not allow the
user to override the RTS-for-aggregation setting
from sysfs any more.

Finally, also clean up the way all these flags get
set in the driver and move everything into the
device-specific functions.

Cc: stable@kernel.org [2.6.35]
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Accesses to "wdev->current_bss" must be
locked with the wdev lock, which action
frame transmission is missing.

Cc: stable@kernel.org [2.6.33+]
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

  CC [M]  drivers/net/wireless/libertas/cfg.o
/home/greearb/git/wireless-testing/drivers/net/wireless/libertas/cfg.c: In function ‘lbs_scan_worker’:
/home/greearb/git/wireless-testing/drivers/net/wireless/libertas/cfg.c:722: error: ‘TASK_NORMAL’ undeclared (first use in this function)
/home/greearb/git/wireless-testing/drivers/net/wireless/libertas/cfg.c:722: error: (Each undeclared identifier is reported only once
/home/greearb/git/wireless-testing/drivers/net/wireless/libertas/cfg.c:722: error: for each function it appears in.)
/home/greearb/git/wireless-testing/drivers/net/wireless/libertas/cfg.c: In function ‘lbs_cfg_connect’:
/home/greearb/git/wireless-testing/drivers/net/wireless/libertas/cfg.c:1267: error: ‘TASK_INTERRUPTIBLE’ undeclared (first use in this function)
/home/greearb/git/wireless-testing/drivers/net/wireless/libertas/cfg.c:1267: error: implicit declaration of function ‘signal_pending’
/home/greearb/git/wireless-testing/drivers/net/wireless/libertas/cfg.c:1267: error: implicit declaration of function ‘schedule_timeout’
Reported-by: Ben Greear <greearb@candelatech.com>
Reported-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

This reverts commit 5f7aebd8.

Apparently, that PCI ID data was incorrectly taken from the subsystem
information.  The actual ID matches another already known ID.
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Adds documentation for the igbvf (igb virtual function driver).

  v2:
   - Removed trailing white space
   - Removed Ethtool version info
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Tested-by: Jeff Pieper <jeffrey.e.pieper@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add documentation for the igb networking driver.

  v2:
   - Removed trailing white space
   - Removed Ethtool version info
   - Removed LRO kernel version info
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Tested-by: Jeff Pieper <jeffrey.e.pieper@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Based on patches from Sonny Rao and Milton Miller...

Combined the patches to fix up clean_tx_irq and clean_rx_irq.

The PowerPC architecture does not require loads to independent bytes
to be ordered without adding an explicit barrier.

In ixgbe_clean_rx_irq we load the status bit then load the packet data.
With packet split disabled if these loads go out of order we get a
stale packet, but we will notice the bad sequence numbers and drop it.

The problem occurs with packet split enabled where the TCP/IP header
and data are in different descriptors. If the reads go out of order
we may have data that doesn't match the TCP/IP header. Since we use
hardware checksumming this bad data is never verified and it makes it
all the way to the application.

This bug was found during stress testing and adding this barrier has
been shown to fix it.  The bug can manifest as a data integrity issue
(bad payload data) or as a BUG in skb_pull().

This was a nasty bug to hunt down, if people agree with the fix I think
it's a candidate for stable.

Previously Submitted to e1000-devel only for ixgbe

http://marc.info/?l=e1000-devel&m=126593062701537&w=3

We've now seen this problem hit with other device drivers (e1000e mostly)
So I'm resubmitting with fixes for other Intel Device Drivers with
similar issues.

CC: Milton Miller <miltonm@bga.com>
CC: Anton Blanchard <anton@samba.org>
CC: Sonny Rao <sonnyrao@us.ibm.com>
CC: stable <stable@kernel.org>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Building ixgbe without DCB_CONFIG and FCOE_CONFIG will cause
a build error.  This resolves the build error by wrapping
the fcoe.up in CONFIG_IXGBE_DCB ifdefs.

Also frames were being priority VLAN tagged even without DCB
enabled.  This fixes this so that 8021Q priority tags are
only added with DCB actually enabled.
Reported-by: divya <dipraksh@linux.vnet.ibm.com>
Reported-by: Jon Mason <jon.mason@exar.com>
Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Amit Kumar Salecha <amit.salecha@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

I moved the kfree(cb) below the dereferences.
Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

We should unlock here.  This is the only place where we return from the
function with the lock held.  The caller isn't expecting it.
Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

We were seeing faults in the solos-pci receive tasklet when packets
arrived for a VCC which was currently being closed:

[18842.727906] EIP: [<e082f490>] br2684_push+0x19/0x234 [br2684] SS:ESP 0068:dfb89d14 

[18845.090712] [<c13ecff3>] ? do_page_fault+0x0/0x2e1 
[18845.120042] [<e082f490>] ? br2684_push+0x19/0x234 [br2684] 
[18845.153530] [<e084fa13>] solos_bh+0x28b/0x7c8 [solos_pci] 
[18845.186488] [<e084f711>] ? solos_irq+0x2d/0x51 [solos_pci] 
[18845.219960] [<c100387b>] ? handle_irq+0x3b/0x48 
[18845.247732] [<c10265cb>] ? irq_exit+0x34/0x57 
[18845.274437] [<c1025720>] tasklet_action+0x42/0x69 
[18845.303247] [<c102643f>] __do_softirq+0x8e/0x129 
[18845.331540] [<c10264ff>] do_softirq+0x25/0x2a 
[18845.358274] [<c102664c>] _local_bh_enable_ip+0x5e/0x6a 
[18845.389677] [<c102666d>] local_bh_enable+0xb/0xe 
[18845.417944] [<e08490a8>] ppp_unregister_channel+0x32/0xbb [ppp_generic] 
[18845.458193] [<e08731ad>] pppox_unbind_sock+0x18/0x1f [pppox] 

This patch uses an RCU-inspired approach to fix it. In the RX tasklet's
find_vcc() function we first refuse to use a VCC which already has the
ATM_VF_READY bit cleared. And in the VCC close function, we synchronise
with the tasklet to ensure that it can't still be using the VCC before
we continue and allow the VCC to be destroyed.
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Tested-by: Nathan Williams <nathan@traverse.com.au>
Cc: stable@kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>

Since there was added ->tcf_chain() method without ->bind_tcf() to
sch_sfq class options, there is oops when a filter is added with
the classid parameter.

Fixes commit 7d2681a6
netdev thread: null pointer at cls_api.c
Signed-off-by: Jarek Poplawski <jarkao2@gmail.com>
Reported-by: Franchoze Eric <franchoze@yandex.ru>
Signed-off-by: David S. Miller <davem@davemloft.net>

Although netif_rx() isn't expected to be called in process context with
preemption enabled, it'd better handle this case. And this is why get_cpu()
is used in the non-RPS #ifdef branch. If tree RCU is selected,
rcu_read_lock() won't disable preemption, so preempt_disable() should be
called explictly.
Signed-off-by: Changli Gao <xiaosuo@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

tcp_parse_md5sig_option doesn't check md5sig option (TCPOPT_MD5SIG)
length, but tcp_v[46]_inbound_md5_hash assume that it's at least 16
bytes long.
Signed-off-by: Dmitry Popov <dp@highloadlab.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

spin_is_locked() can return zero on some (UP?)
configurations because locks don't exist, and
that causes an endless amount of warnings. Use
lockdep_assert_held() instead, which has two
advantages:
 1) it verifies the current task is holding
    the lock or mutex
 2) it compiles away completely when lockdep
    is not enabled

Cc: stable@kernel.org [2.6.34+, maybe only parts of patch]
Reported-by: Thomas Meyer <thomas@m3y3r.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

The TX tracing code copies with the wrong length,
which will typically copy too little data. Fix
this by using the correct length variable.

Cc: stable@kernel.org [2.6.32+]
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

The main motivation of this patch changing strcpy() to strlcpy().
We strcpy() to copy a 48 byte buffers into a 49 byte buffers.  So at
best the last byte has leaked information, or maybe there is an
overflow?  Anyway, this patch closes the information leaks by zeroing
the memory and the calls to strlcpy() prevent overflows.
Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The netpoll_rx_on() check in __napi_gro_receive() skips part of the
"common" GRO_NORMAL path, especially "pull:" in dev_gro_receive(),
where at least eth header should be copied for entirely paged skbs.
Signed-off-by: Jarek Poplawski <jarkao2@gmail.com>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

With turned on hspa modem (Dell 5530 internal card) and activated usb auto
suspend, my system gets up 100 "usbnet_resume has delayed data" per
minute. I didnt noticed any pathological behaviour, so just drop
this message. if any objections, please at least change it to _DEBUG.
Signed-off-by: Christian Samsel <christian.samsel@rwth-aachen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>