pandora-kernel.git
11 years agotcp: xps: fix reordering issues
Eric Dumazet [Thu, 23 May 2013 07:44:20 +0000 (07:44 +0000)]
tcp: xps: fix reordering issues

commit 3853b5841c01a ("xps: Improvements in TX queue selection")
introduced ooo_okay flag, but the condition to set it is slightly wrong.

In our traces, we have seen ACK packets being received out of order,
and RST packets sent in response.

We should test if we have any packets still in host queue.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Tom Herbert <therbert@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet: Revert unused variable changes.
David S. Miller [Thu, 23 May 2013 19:15:32 +0000 (12:15 -0700)]
net: Revert unused variable changes.

This reverts commits:

c573972c111eb4c6b3f3250ad71e7c75cc799833
1a5904342c7380ceddd61c0b37544d752d0b1433
da2e2c214953f37c2a6be20226537ca5a329724c

They were meant for net-next not net.

Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoxfrm: properly handle invalid states as an error
Timo Teräs [Wed, 22 May 2013 01:40:47 +0000 (01:40 +0000)]
xfrm: properly handle invalid states as an error

The error exit path needs err explicitly set. Otherwise it
returns success and the only caller, xfrm_output_resume(),
would oops in skb_dst(skb)->ops derefence as skb_dst(skb) is
NULL.

Bug introduced in commit bb65a9cb (xfrm: removes a superfluous
check and add a statistic).

Signed-off-by: Timo Teräs <timo.teras@iki.fi>
Cc: Li RongQing <roy.qing.li@gmail.com>
Cc: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agovirtio_net: enable napi for all possible queues during open
Jason Wang [Tue, 21 May 2013 20:03:58 +0000 (20:03 +0000)]
virtio_net: enable napi for all possible queues during open

Commit 55257d72bd1c51f25106350f4983ec19f62ed1fa (virtio-net: fill only rx
queues which are being used) only does the napi enabling during open for
curr_queue_pairs. This will break multiqueue receiving since napi of new queues
were still disabled after changing the number of queues.

This patch fixes this by enabling napi for all possible queues during open.

Cc: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Acked-by: Rusty Russell <rusty@rustcorp.com.au>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agotcp: bug fix in proportional rate reduction.
Nandita Dukkipati [Tue, 21 May 2013 15:12:07 +0000 (15:12 +0000)]
tcp: bug fix in proportional rate reduction.

This patch is a fix for a bug triggering newly_acked_sacked < 0
in tcp_ack(.).

The bug is triggered by sacked_out decreasing relative to prior_sacked,
but packets_out remaining the same as pior_packets. This is because the
snapshot of prior_packets is taken after tcp_sacktag_write_queue() while
prior_sacked is captured before tcp_sacktag_write_queue(). The problem
is: tcp_sacktag_write_queue (tcp_match_skb_to_sack() -> tcp_fragment)
adjusts the pcount for packets_out and sacked_out (MSS change or other
reason). As a result, this delta in pcount is reflected in
(prior_sacked - sacked_out) but not in (prior_packets - packets_out).

This patch does the following:
1) initializes prior_packets at the start of tcp_ack() so as to
capture the delta in packets_out created by tcp_fragment.
2) introduces a new "previous_packets_out" variable that snapshots
packets_out right before tcp_clean_rtx_queue, so pkts_acked can be
correctly computed as before.
3) Computes pkts_acked using previous_packets_out, and computes
newly_acked_sacked using prior_packets.

Signed-off-by: Nandita Dukkipati <nanditad@google.com>
Acked-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoMerge tag 'batman-adv-fix-for-davem' of git://git.open-mesh.org/linux-merge
David S. Miller [Thu, 23 May 2013 07:08:36 +0000 (00:08 -0700)]
Merge tag 'batman-adv-fix-for-davem' of git://git.open-mesh.org/linux-merge

Included change:
- fix double free in case of failure during mesh initialisation

Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet: ethernet: sun: drop unused variable
Emilio López [Wed, 22 May 2013 13:57:37 +0000 (13:57 +0000)]
net: ethernet: sun: drop unused variable

Commit bfd428d ("net: ethernet: sun: initialize variables directly")
dropped the only loop that was using i but did not remove the actual
variable, therefore causing a warning when building. This patch drops
the now redundant line.

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Emilio López <emilio@elopez.com.ar>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet: ethernet: korina: drop unused variable
Emilio López [Wed, 22 May 2013 13:57:36 +0000 (13:57 +0000)]
net: ethernet: korina: drop unused variable

Commit e998fd4 ("net: ethernet: korina: initialize variables directly")
dropped the only loop that was using i but did not remove the actual
variable, therefore causing a warning when building. This patch drops
the now redundant line.

Signed-off-by: Emilio López <emilio@elopez.com.ar>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet: ethernet: apple: drop unused variable
Emilio López [Wed, 22 May 2013 13:57:35 +0000 (13:57 +0000)]
net: ethernet: apple: drop unused variable

Commit 3b0aaef ("net: ethernet: apple: initialize variables directly")
dropped the only loop that was using i but did not remove the actual
variable, therefore causing a warning when building. This patch drops
the now redundant line.

Signed-off-by: Emilio López <emilio@elopez.com.ar>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoqmi_wwan: Added support for Cinterion's PLxx WWAN Interface
Hans-Christoph Schemmel [Tue, 21 May 2013 02:07:17 +0000 (02:07 +0000)]
qmi_wwan: Added support for Cinterion's PLxx WWAN Interface

Added support for Cinterion's PLxx WWAN Interface by adding QMI_FIXED_INTF with
Cinterion's Vendor ID as well as Product ID and WWAN Interface Number.

Signed-off-by: Hans-Christoph Schemmel <hans-christoph.schemmel@gemalto.com>
Signed-off-by: Christian Schmiedl <christian.schmiedl@gemalto.com>
Acked-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoMerge branch 'perf_net_dropmonitor'
David S. Miller [Wed, 22 May 2013 22:10:25 +0000 (15:10 -0700)]
Merge branch 'perf_net_dropmonitor'

Ben Hutchings says:

====================
Somewhat surprisingly, the net_dropmonitor reporting script doesn't work
at all.  This series fixes it and then makes it slightly more efficient.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
11 years agoperf: net_dropmonitor: Remove progress indicator
Ben Hutchings [Mon, 20 May 2013 14:45:57 +0000 (14:45 +0000)]
perf: net_dropmonitor: Remove progress indicator

We can read /proc/kallsyms in a fraction of a second, so why waste
a further fraction of a second showing progress?

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoperf: net_dropmonitor: Use bisection in symbol lookup
Ben Hutchings [Mon, 20 May 2013 14:45:42 +0000 (14:45 +0000)]
perf: net_dropmonitor: Use bisection in symbol lookup

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoperf: net_dropmonitor: Do not assume ordering of dictionaries
Ben Hutchings [Mon, 20 May 2013 14:45:34 +0000 (14:45 +0000)]
perf: net_dropmonitor: Do not assume ordering of dictionaries

The sort order of dictionaries in Python is undocumented.  Use
tuples instead, which are documented to be lexically ordered.

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoperf: net_dropmonitor: Fix symbol-relative addresses
Ben Hutchings [Mon, 20 May 2013 14:45:26 +0000 (14:45 +0000)]
perf: net_dropmonitor: Fix symbol-relative addresses

The comparison between traced and symbol addresses is backwards: if
the traced address doesn't exactly match a symbol (which we don't
expect it to), we'll show the next symbol and the offset to it,
whereas we should show the previous symbol and the offset from it.

Cc: stable@vger.kernel.org
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoperf: net_dropmonitor: Fix trace parameter order
Ben Hutchings [Mon, 20 May 2013 14:44:43 +0000 (14:44 +0000)]
perf: net_dropmonitor: Fix trace parameter order

This works much better if we don't treat protocol numbers as addresses.

Cc: stable@vger.kernel.org
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet: fec: use a more proper compatible string for MVF type device
Shawn Guo [Sun, 19 May 2013 04:38:46 +0000 (04:38 +0000)]
net: fec: use a more proper compatible string for MVF type device

MVF is a family while MVF600 is a particular SoC in the family.  We
generally prefer to use SoC rather than family name in compatible string
to define a particular type of fec device.  And this is how fec_dt_ids
works for all those IMX fec variants.  Let's change mvf to mvf600 to
have it work in the same way.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoqlcnic: Fix updating netdev->features
Shahed Shaikh [Tue, 21 May 2013 09:27:01 +0000 (09:27 +0000)]
qlcnic: Fix updating netdev->features

o After change in EPORT features of 82xx adapter, netdev->features needs to
  be updated to reflect EPORT feature updates but driver was manipulating
  netdev->features at wrong place.
o This patch uses netdev_update_features() and .ndo_fix_features() to
  update netdev->features properly.

Signed-off-by: Shahed Shaikh <shahed.shaikh@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoqlcnic: remove netdev->trans_start updates within the driver
Sony Chacko [Tue, 21 May 2013 09:27:00 +0000 (09:27 +0000)]
qlcnic: remove netdev->trans_start updates within the driver

Code is removed because netdev->trans_start updates made by the driver
will be ignored by the kernel.

Signed-off-by: Sony Chacko <sony.chacko@qlogic.com>
Signed-off-by: Shahed Shaikh <shahed.shaikh@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoqlcnic: Return proper error codes from probe failure paths
Sony Chacko [Tue, 21 May 2013 09:26:59 +0000 (09:26 +0000)]
qlcnic: Return proper error codes from probe failure paths

Fix error paths in probe to assign proper error codes to probe return value.

Signed-off-by: Sony Chacko <sony.chacko@qlogic.com>
Signed-off-by: Shahed Shaikh <shahed.shaikh@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agotg3: Update version to 3.132
Nithin Sujir [Tue, 21 May 2013 12:57:33 +0000 (12:57 +0000)]
tg3: Update version to 3.132

Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agotg3: Ensure boot code has completed initialization before accessing hardware
Nithin Sujir [Tue, 21 May 2013 12:57:32 +0000 (12:57 +0000)]
tg3: Ensure boot code has completed initialization before accessing hardware

After resetting the device, the driver waits for a signature to be
updated to know that firmware has completed initialization. However, the
call to tg3_poll_fw() is being done too late and we're writing to the
GRC_MODE register before it has completely initialized, causing
contention with firmware.  This logic has existed since day one but is
causing PCIE link to go down randomly at startup on one platform once
every few hundred reboots.

Move the tg3_poll_fw() up to before we write to the GRC_MODE register
after reset.

Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobatman-adv: Avoid double freeing of bat_counters
Martin Hundebøll [Wed, 17 Apr 2013 19:13:16 +0000 (21:13 +0200)]
batman-adv: Avoid double freeing of bat_counters

On errors in batadv_mesh_init(), bat_counters will be freed in both
batadv_mesh_free() and batadv_softif_init_late(). This patch fixes this
by returning earlier from batadv_softif_init_late() in case of errors in
batadv_mesh_init() and by setting bat_counters to NULL after freeing.

Signed-off-by: Martin Hundebøll <martin@hundeboll.net>
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
Signed-off-by: Antonio Quartulli <ordex@autistici.org>
11 years agoMerge branch 'for-davem' of git://git.kernel.org/pub/scm/linux/kernel/git/linville...
David S. Miller [Mon, 20 May 2013 21:05:22 +0000 (14:05 -0700)]
Merge branch 'for-davem' of git://git./linux/kernel/git/linville/wireless

John W. Linville says:

====================
This pull request is intended for the 3.10 series.  It contains a
variety of fixes for problems discovered during the merge window and
after 3.10-rc1.

For the mac80211 bits, Johannes says the following:

"This is what I have:
 * a patch from Felix to fix RCU usage in his rate table code
 * a patch from Ilan to add the wdev id to some notifications so they can
   actually be used by userspace
 * Sasha Levin found an issue in how hwsim handles devices
 * a fix for a bug in the wiphy_register() error path that's been there forever
 * three fixes for WoWLAN
 * AP mode frame matching was erroneously giving frames to all virtual AP
   interfaces (reported by Jouni)
 * a fix for HT handling in my CSA changes, found by Sujith
 * a fix for some locking simplifications gone wrong
 * Ben Greear found more cfg80211/mac80211 state confusion
 * and a fix for another bug found by Jouni: local state changes need to be
   reported by mac80211 to cfg80211 so it disconnects properly."

And for the iwlwifi bits, he says:

"I have fixes for a firmware crash during resume, multicast RX,
aggregation and a workaround for a firmware scanning bug."

Along with those...

Albert Pool adds a USB ID to the rtl8192cu driver.

Arend van Spriel restores a driver option support flag that had been
removed from 3.9 due to a bug in that version of the driver.

Felix Fietkau fixes a trio of ath9k issues with a series of small
patches.

Geert Uytterhoeven provides a Kconfig fix for ath9k (which you also
merged, so it isn't in the diff here).

Larry Finger gives us a fix for a build warning on big-endian systems
for rtlwifi.

Rafał Miłecki adds some core IDs to the bcma driver.

Sujith Manoharan fixes a module unloading crash in ath9k, and corrects
some calibration settings for AR9485.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
11 years ago8139cp: reset BQL when ring tx ring cleared
stephen hemminger [Mon, 20 May 2013 06:54:43 +0000 (06:54 +0000)]
8139cp: reset BQL when ring tx ring cleared

This patch cures transmit timeout's with DHCP observed
while running under KVM. When the transmit ring is cleaned out,
the Byte Queue Limit values need to be reset.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agodrivers: net: usb: rtl8150: concurrent URB bugfix
Petko Manolov [Sun, 19 May 2013 23:08:47 +0000 (23:08 +0000)]
drivers: net: usb: rtl8150: concurrent URB bugfix

This patch fixes a potential race with concurrently running asynchronous
write requests.  The values for device's RX control register are now
stored in dynamically allocated buffers so each URB submission has it's
own copy.  Doing it the old way is data clobbering prone.

This patch is against latest 'net' tree.

Signed-off-by: Petko Manolov <petkan@nucleusys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoisdn/kcapi: fix a small underflow
Dan Carpenter [Sun, 19 May 2013 08:36:36 +0000 (08:36 +0000)]
isdn/kcapi: fix a small underflow

In get_capi_ctr_by_nr() and get_capi_appl_by_nr() the parameter comes
from skb->data.  The current code can underflow to one space before the
start of the array.

The sanity check isn't needed in __get_capi_appl_by_nr() but I changed
it to match the others.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wirel...
John W. Linville [Mon, 20 May 2013 19:19:01 +0000 (15:19 -0400)]
Merge branch 'master' of git://git./linux/kernel/git/linville/wireless into for-davem

11 years agobnx2x: Fix GSO for 57710/57711 chips
Yuval Mintz [Sun, 19 May 2013 04:41:01 +0000 (04:41 +0000)]
bnx2x: Fix GSO for 57710/57711 chips

Starting with commit 91226790bbe2dbfbba48dd79d49f2b38ef10eb97
`bnx2x: use FW 7.8.17', the bnx2x driver no longer requests the FW to perform
IP checksums for IPv4 packets.

This behaviour needs to be revised for 57710/57711 chips -
when using GSO, if the driver will not set the IP checksum flag then packets
will be transmitted by the chip without a valid IP checksum, resulting in
a drop of all such packets on the receiver-side.

Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoip_gre: fix a possible crash in ipgre_err()
Eric Dumazet [Sat, 18 May 2013 08:36:03 +0000 (08:36 +0000)]
ip_gre: fix a possible crash in ipgre_err()

Another fix needed in ipgre_err(), as parse_gre_header() might change
skb->head.

Bug added in commit c54419321455 (GRE: Refactor GRE tunneling code.)

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Pravin B Shelar <pshelar@nicira.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agor8169: fix offloaded tx checksum for small packets.
françois romieu [Sat, 18 May 2013 01:24:46 +0000 (01:24 +0000)]
r8169: fix offloaded tx checksum for small packets.

8168evl offloaded checksums are wrong since commit
e5195c1f31f399289347e043d6abf3ffa80f0005 ("r8169: fix 8168evl frame padding.")
pads small packets to 60 bytes (without ethernet checksum). Typical symptoms
appear as UDP checksums which are wrong by the count of added bytes.

It isn't worth compensating. Let the driver checksum.

Due to the skb length changes, TSO code is moved before the Tx descriptor gets
written.

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Tested-by: Holger Hoffstätte <holger.hoffstaette@googlemail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobonding: fix multiple 3ad mode sysfs race conditions
nikolay@redhat.com [Sat, 18 May 2013 01:18:31 +0000 (01:18 +0000)]
bonding: fix multiple 3ad mode sysfs race conditions

When bond_3ad_get_active_agg_info() is used in all show_ad_ functions
it is not protected against slave manipulation and since it walks over
the slaves and uses them, this can easily result in NULL pointer
dereference or use of freed memory. Both the new wrapper and the
internal function are exported to the bonding as they're needed in
different places.

Signed-off-by: Nikolay Aleksandrov <nikolay@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobonding: arp_ip_count and arp_targets can be wrong
nikolay@redhat.com [Sat, 18 May 2013 01:18:30 +0000 (01:18 +0000)]
bonding: arp_ip_count and arp_targets can be wrong

When getting arp_ip_targets if we encounter a bad IP, arp_ip_count still
gets increased and all the targets after the wrong one will not be probed
if arp_interval is enabled after that (unless a new IP target is added
through sysfs) because of the zero entry, in this case reading
arp_ip_target through sysfs will show valid targets even if there's a
zero entry.
Example: 1.2.3.4,4.5.6.7,blah,5.6.7.8
When retrieving the list from arp_ip_target the output would be:
1.2.3.4,4.5.6.7,5.6.7.8
but there will be a 0 entry between 4.5.6.7 and 5.6.7.8. If arp_interval
is enabled after that 5.6.7.8 will never be checked because of that.

Signed-off-by: Nikolay Aleksandrov <nikolay@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobonding: replace %x with %pI4 for IPv4 addresses
nikolay@redhat.com [Sat, 18 May 2013 01:18:29 +0000 (01:18 +0000)]
bonding: replace %x with %pI4 for IPv4 addresses

There're few pr_debug() places that can provide the IPv4 address in
dotted decimal format instead which is more helpful.

Signed-off-by: Nikolay Aleksandrov <nikolay@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobonding: fix set mode race conditions
nikolay@redhat.com [Sat, 18 May 2013 01:18:28 +0000 (01:18 +0000)]
bonding: fix set mode race conditions

Changing the mode without any locking can result in multiple races (e.g.
upping a bond, enslaving/releasing). Depending on which race is hit the
impact can vary from incosistent bond state to kernel crash.
Use RTNL to synchronize the mode setting with the dangerous races.

Signed-off-by: Nikolay Aleksandrov <nikolay@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet: irda: using kzalloc() instead of kmalloc() to avoid strncpy() issue.
Chen Gang [Thu, 16 May 2013 23:13:04 +0000 (23:13 +0000)]
net: irda: using kzalloc() instead of kmalloc() to avoid strncpy() issue.

'discovery->data.info' length is 22, NICKNAME_MAX_LEN is 21, so the
strncpy() will always left the last byte of 'discovery->data.info'
uninitialized.

When 'text' length is longer than 21 (NICKNAME_MAX_LEN), if still left
the last byte of 'discovery->data.info' uninitialized, the next
strlen() will cause issue.

Also 'discovery->data' is 'struct irda_device_info' which defined in
"include/uapi/...", it may copy to user mode, so need whole initialized.

All together, need use kzalloc() instead of kmalloc() to initialize all
members firstly.

Signed-off-by: Chen Gang <gang.chen@asianux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonetlabel: improve domain mapping validation
Paul Moore [Fri, 17 May 2013 09:08:50 +0000 (09:08 +0000)]
netlabel: improve domain mapping validation

The net/netlabel/netlabel_domainhash.c:netlbl_domhsh_add() function
does not properly validate new domain hash entries resulting in
potential problems when an administrator attempts to add an invalid
entry.  One such problem, as reported by Vlad Halilov, is a kernel
BUG (found in netlabel_domainhash.c:netlbl_domhsh_audit_add()) when
adding an IPv6 outbound mapping with a CIPSO configuration.

This patch corrects this problem by adding the necessary validation
code to netlbl_domhsh_add() via the newly created
netlbl_domhsh_validate() function.

Ideally this patch should also be pushed to the currently active
-stable trees.

Reported-by: Vlad Halilov <vlad.halilov@gmail.com>
Signed-off-by: Paul Moore <pmoore@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoipv6: fix possible crashes in ip6_cork_release()
Eric Dumazet [Fri, 17 May 2013 04:53:13 +0000 (04:53 +0000)]
ipv6: fix possible crashes in ip6_cork_release()

commit 0178b695fd6b4 ("ipv6: Copy cork options in ip6_append_data")
added some code duplication and bad error recovery, leading to potential
crash in ip6_cork_release() as kfree() could be called with garbage.

use kzalloc() to make sure this wont happen.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>
Cc: Neal Cardwell <ncardwell@google.com>
11 years agovxlan: Update vxlan fdb 'used' field after each usage
Sridhar Samudrala [Fri, 17 May 2013 06:39:07 +0000 (06:39 +0000)]
vxlan: Update vxlan fdb 'used' field after each usage

Fix some instances where vxlan fdb 'used' field is not updated after the entry
is used.

v2: rename vxlan_find_mac() as __vxlan_find_mac() and create a new vxlan_find_mac()
that also updates ->used field.

Signed-off-by: Sridhar Samudrala <sri@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agogianfar: add missing iounmap() on error in gianfar_ptp_probe()
Wei Yongjun [Thu, 16 May 2013 22:25:34 +0000 (22:25 +0000)]
gianfar: add missing iounmap() on error in gianfar_ptp_probe()

Add the missing iounmap() before return from gianfar_ptp_probe()
in the error handling case.

Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoath9k: fix draining aggregation tid buffers
Felix Fietkau [Fri, 17 May 2013 10:58:26 +0000 (12:58 +0200)]
ath9k: fix draining aggregation tid buffers

After a tx attempt, an A-MPDU subframe can still have fi->retries at 0
(if the retry count wasn't incremented due to powersave).
In that case it is still tracked as part of the block ack window, so
when draining the tid queue, its sequence number needs to be cleared
from the pending frame bitmap.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
11 years agoath9k: fix rate handling/reporting
Felix Fietkau [Fri, 17 May 2013 10:58:25 +0000 (12:58 +0200)]
ath9k: fix rate handling/reporting

This patch fixes some issues introduced in the rate control API rework.
When not running aggregation, copy bf->rates into info->control.rates
before applying the rate control status to it.
In ath_lookup_rate, the rates need to be pulled from bf->rates, not the
tx info.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
11 years agoath9k: fix aggregation stop/flush handling
Felix Fietkau [Fri, 17 May 2013 10:58:24 +0000 (12:58 +0200)]
ath9k: fix aggregation stop/flush handling

When aggregation stop is requested, don't run the mac80211 aggregation
stop callback yet, while the session is still blocked.
Also, when aggregation flush is requested, don't run the callback at all.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
11 years agoath9k_hw: Enable manual peak calibration for AR9485
Sujith Manoharan [Thu, 16 May 2013 17:17:34 +0000 (22:47 +0530)]
ath9k_hw: Enable manual peak calibration for AR9485

Manual peak calibration is currently enabled only for
AR9462 and AR9565. This is also required for AR9485.
The initvals are also modified to disable HW peak calibration.

Cc: <stable@vger.kernel.org>
Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
11 years agortlwifi: rtl8192cu: Add new USB ID
Albert Pool [Wed, 15 May 2013 15:03:16 +0000 (10:03 -0500)]
rtlwifi: rtl8192cu: Add new USB ID

This adds the USB ID of the On Networks N300MA, clone of Netgear WNA3100M.

Signed-off-by: Albert Pool <albertpool@solcon.nl>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Reported-by: Ana Rey <Anazul77@hotmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
11 years agobrcmfmac: announce P2P_DEVICE support in wiphy structure
Arend van Spriel [Tue, 14 May 2013 18:52:36 +0000 (20:52 +0200)]
brcmfmac: announce P2P_DEVICE support in wiphy structure

P2P_DEVICE support was removed from brcmfmac for v3.9 kernel with
the commit below:

commit 1527c343c12f3a2aae532aa881d12c6fbf8749f4
Author: Arend van Spriel <arend@broadcom.com>
Date:   Thu Apr 4 12:10:11 2013 +0200

    brcmfmac: remove advertising P2P device support

However, it got merged into wireless-next. But for 3.10 brcmfmac does
support P2P device. Putting it back with this commit.

Reviewed-by: Hante Meuleman <meuleman@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
11 years agortlwifi: rtl8188ee: Fix warning when building on big-endian systems
Larry Finger [Mon, 13 May 2013 02:43:46 +0000 (21:43 -0500)]
rtlwifi: rtl8188ee: Fix warning when building on big-endian systems

In http://lkml.indiana.edu/hypermail/linux/kernel/1305.1/index.html,
Geert Uytterhoeven reports a new warning when building 3.10-rc1 in
this driver. This is caused by using a "#if" test to see if __LITTLE_ENDIAN
is set, which fails for all big-endian systems. Change to "ifdef".

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
11 years agoath9k: Fix crash on module unload
Sujith Manoharan [Fri, 10 May 2013 13:11:06 +0000 (18:41 +0530)]
ath9k: Fix crash on module unload

Make sure that any open relayfs files are closed before
unregistering with mac80211, otherwise this crash is seen:

[ 1331.097846] BUG: unable to handle kernel paging request at 6b6b6b8b
[ 1331.098170] IP: [<c063d0d6>] debugfs_remove+0x26/0x80
[ 1331.098170] *pdpt = 000000002f9aa001 *pde = 0000000000000000
[ 1331.098170] Oops: 0000 [#1] PREEMPT SMP
[ 1331.098170] Modules linked in: iptable_raw xt_CT nf_conntrack_ipv4 nf_defrag]
[ 1331.098170] Pid: 4794, comm: rmmod Tainted: G        WC   3.9.1+ #5 To Be Fi.
[ 1331.098170] EIP: 0060:[<c063d0d6>] EFLAGS: 00010202 CPU: 0
[ 1331.098170] EIP is at debugfs_remove+0x26/0x80
[ 1331.098170] EAX: f2f3acd0 EBX: f2f3acd0 ECX: 00000006 EDX: f8622348
[ 1331.098170] ESI: 6b6b6b6b EDI: 00000001 EBP: ee251e14 ESP: ee251e0c
[ 1331.098170]  DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS: 0068
[ 1331.098170] CR0: 8005003b CR2: 6b6b6b8b CR3: 2e7b7000 CR4: 000007e0
[ 1331.098170] DR0: 00000000 DR1: 00000000 DR2: 00000000 DR3: 00000000
[ 1331.098170] DR6: ffff0ff0 DR7: 00000400
[ 1331.098170] Process rmmod (pid: 4794, ti=ee250000 task=efaa2560 task.ti=ee25)
[ 1331.098170] Stack:
[ 1331.098170]  f241e170 0000000a ee251e1c f861394d ee251e28 c04e3088 f241e170 4
[ 1331.098170]  c04e30fe f45482b0 ee251e54 c04e3187 f25e86b0 ee251e54 f8618748 0
[ 1331.098170]  0000000a 00000001 ee251e68 f860065b f2509e20 f25085a0 f5b6e8a4 8
[ 1331.098170] Call Trace:
[ 1331.098170]  [<f861394d>] remove_buf_file_handler+0xd/0x20 [ath9k]
[ 1331.098170]  [<c04e3088>] relay_remove_buf+0x18/0x30
[ 1331.098170]  [<c04e30fe>] relay_close_buf+0x2e/0x40
[ 1331.098170]  [<c04e3187>] relay_close+0x77/0xf0
[ 1331.098170]  [<f8618748>] ? dpd_exit+0x38/0x40 [ath9k]
[ 1331.098170]  [<f860065b>] ath9k_deinit_softc+0x8b/0xa0 [ath9k]
[ 1331.098170]  [<f86006b8>] ath9k_deinit_device+0x48/0x60 [ath9k]
[ 1331.098170]  [<f86107f1>] ath_pci_remove+0x31/0x50 [ath9k]
[ 1331.098170]  [<c06dbff8>] pci_device_remove+0x38/0xc0
[ 1331.098170]  [<c079daa4>] __device_release_driver+0x64/0xc0
[ 1331.098170]  [<c079db97>] driver_detach+0x97/0xa0
[ 1331.098170]  [<c079cacc>] bus_remove_driver+0x6c/0xe0
[ 1331.098170]  [<c079c197>] ? bus_put+0x17/0x20
[ 1331.098170]  [<c079cae3>] ? bus_remove_driver+0x83/0xe0
[ 1331.098170]  [<c079e709>] driver_unregister+0x49/0x80
[ 1331.098170]  [<c06dc138>] pci_unregister_driver+0x18/0x80
[ 1331.098170]  [<f8610602>] ath_pci_exit+0x12/0x20 [ath9k]
[ 1331.098170]  [<f8619ce0>] ath9k_exit+0x17/0x337 [ath9k]
[ 1331.098170]  [<c09e537d>] ? mutex_unlock+0xd/0x10
[ 1331.098170]  [<c04bd36c>] sys_delete_module+0x17c/0x250
[ 1331.098170]  [<c0540dc4>] ? do_munmap+0x244/0x2d0
[ 1331.098170]  [<c0540e96>] ? vm_munmap+0x46/0x60
[ 1331.098170]  [<c09e8dc4>] ? restore_all+0xf/0xf
[ 1331.098170]  [<c09ebf50>] ? __do_page_fault+0x4c0/0x4c0
[ 1331.098170]  [<c04b18e4>] ? trace_hardirqs_on_caller+0xf4/0x180
[ 1331.098170]  [<c09ef28d>] sysenter_do_call+0x12/0x38
[ 1331.098170] Code: 90 8d 74 26 00 55 89 e5 83 ec 08 89 1c 24 89 74 24 04 3e 82
[ 1331.098170] EIP: [<c063d0d6>] debugfs_remove+0x26/0x80 SS:ESP 0068:ee251e0c
[ 1331.098170] CR2: 000000006b6b6b8b
[ 1331.727971] ---[ end trace b5bb9f2066cef7f9 ]---

Cc: <stable@vger.kernel.org>
Acked-by: Simon Wunderlich <siwu@hrz.tu-chemnitz.de>
Tested-by: Ben Greear <greearb@candelatech.com>
Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
11 years agonet/wireless: ATH9K should depend on HAS_DMA
Geert Uytterhoeven [Thu, 9 May 2013 21:04:47 +0000 (23:04 +0200)]
net/wireless: ATH9K should depend on HAS_DMA

If NO_DMA=y:

drivers/built-in.o: In function `ath9k_beacon_generate':
drivers/net/wireless/ath/ath9k/beacon.c:146: undefined reference to `dma_unmap_single'
drivers/net/wireless/ath/ath9k/beacon.c:174: undefined reference to `dma_map_single'
drivers/net/wireless/ath/ath9k/beacon.c:176: undefined reference to `dma_mapping_error'
drivers/built-in.o: In function `ath9k_beacon_remove_slot':
drivers/net/wireless/ath/ath9k/beacon.c:252: undefined reference to `dma_unmap_single'
drivers/built-in.o: In function `ath_descdma_setup':
drivers/net/wireless/ath/ath9k/init.c:382: undefined reference to `dmam_alloc_coherent'
drivers/built-in.o: In function `ath_edma_get_buffers':
drivers/net/wireless/ath/ath9k/recv.c:616: undefined reference to `dma_sync_single_for_cpu'
drivers/built-in.o: In function `ath_get_next_rx_buf':
drivers/net/wireless/ath/ath9k/recv.c:740: undefined reference to `dma_sync_single_for_cpu'
drivers/built-in.o: In function `ath_rx_edma_cleanup':
drivers/net/wireless/ath/ath9k/recv.c:176: undefined reference to `dma_unmap_single'
drivers/built-in.o: In function `ath_rx_cleanup':
drivers/net/wireless/ath/ath9k/recv.c:340: undefined reference to `dma_unmap_single'
drivers/built-in.o: In function `ath_rx_edma_buf_link':
drivers/net/wireless/ath/ath9k/recv.c:122: undefined reference to `dma_sync_single_for_cpu'
drivers/built-in.o: In function `ath_rx_tasklet':
drivers/net/wireless/ath/ath9k/recv.c:1275: undefined reference to `dma_map_single'
drivers/net/wireless/ath/ath9k/recv.c:1277: undefined reference to `dma_mapping_error'
drivers/net/wireless/ath/ath9k/recv.c:1283: undefined reference to `dma_unmap_single'
drivers/built-in.o: In function `ath_rx_edma_init':
drivers/net/wireless/ath/ath9k/recv.c:226: undefined reference to `dma_map_single'
drivers/net/wireless/ath/ath9k/recv.c:229: undefined reference to `dma_mapping_error'
drivers/built-in.o: In function `ath_rx_init':
drivers/net/wireless/ath/ath9k/recv.c:303: undefined reference to `dma_map_single'
drivers/net/wireless/ath/ath9k/recv.c:306: undefined reference to `dma_mapping_error'
drivers/built-in.o: In function `ath_tx_complete_buf':
drivers/net/wireless/ath/ath9k/xmit.c:2088: undefined reference to `dma_unmap_single'
drivers/built-in.o: In function `ath_txstatus_setup':
drivers/net/wireless/ath/ath9k/xmit.c:2344: undefined reference to `dmam_alloc_coherent'
drivers/built-in.o: In function `ath_tx_set_retry':
drivers/net/wireless/ath/ath9k/xmit.c:307: undefined reference to `dma_sync_single_for_cpu'
drivers/built-in.o: In function `ath_tx_setup_buffer':
drivers/net/wireless/ath/ath9k/xmit.c:1887: undefined reference to `dma_map_single'
drivers/net/wireless/ath/ath9k/xmit.c:1889: undefined reference to `dma_mapping_error'

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Luis R. Rodriguez <mcgrof@qca.qualcomm.com>
Cc: John W. Linville <linville@tuxdriver.com>
Cc: linux-wireless@vger.kernel.org
Cc: netdev@vger.kernel.org
Signed-off-by: John W. Linville <linville@tuxdriver.com>
11 years agobcma: add more core IDs
Rafał Miłecki [Thu, 9 May 2013 19:24:24 +0000 (21:24 +0200)]
bcma: add more core IDs

PCIe and ARM CR4 cores were found on 14e4:43b1 AKA BCM4352.

Reported-by: Gabriel Thörnblad <gabriel@thornblad.com>
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
11 years agoiwlegacy: remove inline marking of EXPORT_SYMBOL functions
Denis Efremov [Thu, 9 May 2013 09:19:39 +0000 (13:19 +0400)]
iwlegacy: remove inline marking of EXPORT_SYMBOL functions

EXPORT_SYMBOL and inline directives are contradictory to each other.
The patch fixes this inconsistency.

Found by Linux Driver Verification project (linuxtesting.org).

Signed-off-by: Denis Efremov <yefremov.denis@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
11 years agoMerge branch 'for-john' of git://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi...
John W. Linville [Fri, 17 May 2013 18:22:17 +0000 (14:22 -0400)]
Merge branch 'for-john' of git://git./linux/kernel/git/iwlwifi/iwlwifi-fixes

11 years agoMerge tag 'batman-adv-fix-for-davem' of git://git.open-mesh.org/linux-merge
David S. Miller [Fri, 17 May 2013 08:32:20 +0000 (01:32 -0700)]
Merge tag 'batman-adv-fix-for-davem' of git://git.open-mesh.org/linux-merge

Included changes:
- fix global protection fault by avoiding double call_rcu

Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobatman-adv: Fix rcu_barrier() miss due to double call_rcu() in TT code
Linus Lüssing [Mon, 15 Apr 2013 13:43:29 +0000 (21:43 +0800)]
batman-adv: Fix rcu_barrier() miss due to double call_rcu() in TT code

rcu_barrier() only waits for the currently scheduled rcu functions
to finish - it won't wait for any function scheduled via another
call_rcu() within an rcu scheduled function.

Unfortunately our batadv_tt_orig_list_entry_free_ref() does just that,
via a batadv_orig_node_free_ref() call, leading to our rcu_barrier()
call potentially missing such a batadv_orig_node_free_ref().

This patch fixes this issue by calling the batadv_orig_node_free_rcu()
directly from the rcu callback, removing the unnecessary, additional
call_rcu() layer here.

Signed-off-by: Linus Lüssing <linus.luessing@web.de>
Signed-off-by: Marek Lindner <lindner_marek@yahoo.de>
Acked-by: Antonio Quartulli <ordex@autistici.org>
11 years agobonding: allow TSO being set on bonding master
Eric Dumazet [Thu, 16 May 2013 07:34:53 +0000 (07:34 +0000)]
bonding: allow TSO being set on bonding master

In some situations, we need to disable TSO on bonding slaves.

bonding device automatically unset TSO in bond_fix_features(), and
performance is not good because :

1) We consume more cpu cycles.

2) GSO segmentation has some bugs leading to out of order TCP packets
if this segmentation is done before virtual device. This particular
problem will be addressed in a separate patch.

This patch allows TSO being set/unset on the bonding master,
so that GSO segmentation is done after bonding layer.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Michał Mirosław <mirqus@gmail.com>
Cc: Jay Vosburgh <fubar@us.ibm.com>
Cc: Andy Gospodarek <andy@greyhouse.net>
Cc: Maciej Żenczykowski <maze@google.com>
Cc: Tom Herbert <therbert@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoNET: mv643xx_eth: avoid lockdep dump on interface down
Russell King [Thu, 16 May 2013 06:31:42 +0000 (06:31 +0000)]
NET: mv643xx_eth: avoid lockdep dump on interface down

When the interface is shutdown, the mv643xx_eth driver hits the following
lockdep dump:

=================================
[ INFO: inconsistent lock state ]
3.8.0+ #303 Not tainted
---------------------------------
inconsistent {IN-SOFTIRQ-W} -> {SOFTIRQ-ON-W} usage.
NetworkManager/3449 [HC0[0]:SC0[0]:HE1:SE1] takes:
 (_xmit_ETHER#2){+.?...}, at: [<c02828e4>] txq_reclaim+0x60/0x230
{IN-SOFTIRQ-W} state was registered at:
  [<c007e93c>] mark_irqflags+0xf8/0x1c4
  [<c007ee60>] __lock_acquire+0x458/0x9a4
  [<c007f8b0>] lock_acquire+0x60/0x74
  [<c03ea914>] _raw_spin_lock+0x40/0x50
  [<c0334040>] sch_direct_xmit+0xa4/0x2e4
  [<c0320880>] dev_queue_xmit+0x174/0x508
  [<c03953b0>] ip6_finish_output2+0xd0/0x3c4
  [<c03b15bc>] mld_sendpack+0x190/0x368
  [<c03b3204>] mld_ifc_timer_expire+0xc/0x58
  [<c005133c>] call_timer_fn+0x6c/0xe0
  [<c0051588>] run_timer_softirq+0x1d8/0x210
  [<c004c004>] __do_softirq+0xe0/0x1b4
  [<c004c448>] irq_exit+0x64/0x6c
  [<c000f1e0>] handle_IRQ+0x34/0x84
  [<c000e0d0>] __irq_usr+0x30/0x80
irq event stamp: 160603
hardirqs last  enabled at (160603): [<c00c736c>] kfree+0xa8/0xe8
hardirqs last disabled at (160602): [<c00c72e0>] kfree+0x1c/0xe8
softirqs last  enabled at (160304): [<c028260c>] mib_counters_update+0x5ec/0x60c
softirqs last disabled at (160302): [<c03eab8c>] _raw_spin_lock_bh+0x14/0x54

other info that might help us debug this:
 Possible unsafe locking scenario:

       CPU0
       ----
  lock(_xmit_ETHER#2);
  <Interrupt>
    lock(_xmit_ETHER#2);

 *** DEADLOCK ***

1 lock held by NetworkManager/3449:
 #0:  (rtnl_mutex){+.+.+.}, at: [<c032e664>] rtnetlink_rcv+0xc/0x24

stack backtrace:
[<c0013e34>] (unwind_backtrace+0x0/0xf8) from [<c007e12c>] (print_usage_bug+0x150/0x1d4)
[<c007e12c>] (print_usage_bug+0x150/0x1d4) from [<c007e3f8>] (mark_lock_irq+0x248/0x290)
[<c007e3f8>] (mark_lock_irq+0x248/0x290) from [<c007e598>] (mark_lock+0x158/0x404)
[<c007e598>] (mark_lock+0x158/0x404) from [<c007e97c>] (mark_irqflags+0x138/0x1c4)
[<c007e97c>] (mark_irqflags+0x138/0x1c4) from [<c007ee60>] (__lock_acquire+0x458/0x9a4)
[<c007ee60>] (__lock_acquire+0x458/0x9a4) from [<c007f8b0>] (lock_acquire+0x60/0x74)
[<c007f8b0>] (lock_acquire+0x60/0x74) from [<c03ea914>] (_raw_spin_lock+0x40/0x50)
[<c03ea914>] (_raw_spin_lock+0x40/0x50) from [<c02828e4>] (txq_reclaim+0x60/0x230)
[<c02828e4>] (txq_reclaim+0x60/0x230) from [<c0282ad8>] (txq_deinit+0x24/0xcc)
[<c0282ad8>] (txq_deinit+0x24/0xcc) from [<c0282d28>] (mv643xx_eth_stop+0x1a8/0x1bc)
[<c0282d28>] (mv643xx_eth_stop+0x1a8/0x1bc) from [<c031e314>] (__dev_close_many+0x88/0xcc)
[<c031e314>] (__dev_close_many+0x88/0xcc) from [<c031e380>] (__dev_close+0x28/0x3c)
[<c031e380>] (__dev_close+0x28/0x3c) from [<c0320fa0>] (__dev_change_flags+0x7c/0x134)
[<c0320fa0>] (__dev_change_flags+0x7c/0x134) from [<c03210e0>] (dev_change_flags+0x10/0x48)
[<c03210e0>] (dev_change_flags+0x10/0x48) from [<c032da1c>] (do_setlink+0x1a0/0x730)
[<c032da1c>] (do_setlink+0x1a0/0x730) from [<c032f524>] (rtnl_newlink+0x304/0x4b0)
[<c032f524>] (rtnl_newlink+0x304/0x4b0) from [<c032ef8c>] (rtnetlink_rcv_msg+0x25c/0x2a0)
[<c032ef8c>] (rtnetlink_rcv_msg+0x25c/0x2a0) from [<c03383a0>] (netlink_rcv_skb+0xbc/0xd8)
[<c03383a0>] (netlink_rcv_skb+0xbc/0xd8) from [<c032e674>] (rtnetlink_rcv+0x1c/0x24)
[<c032e674>] (rtnetlink_rcv+0x1c/0x24) from [<c03361d8>] (netlink_unicast_kernel+0x88/0xd4)
[<c03361d8>] (netlink_unicast_kernel+0x88/0xd4) from [<c0337dd0>] (netlink_unicast+0x138/0x180)
[<c0337dd0>] (netlink_unicast+0x138/0x180) from [<c0338020>] (netlink_sendmsg+0x208/0x32c)
[<c0338020>] (netlink_sendmsg+0x208/0x32c) from [<c030ab48>] (sock_sendmsg+0x84/0xa4)
[<c030ab48>] (sock_sendmsg+0x84/0xa4) from [<c030aef4>] (__sys_sendmsg+0x2ac/0x2c4)
[<c030aef4>] (__sys_sendmsg+0x2ac/0x2c4) from [<c030c8ec>] (sys_sendmsg+0x3c/0x68)
[<c030c8ec>] (sys_sendmsg+0x3c/0x68) from [<c000e2e0>] (ret_fast_syscall+0x0/0x3c)

It seems that txq_reclaim() takes the netif tx lock:

        __netif_tx_lock(nq, smp_processor_id());

in a context outside of softirq context, and thus is susceptible to
deadlock should an interrupt occur.

Use __netif_tx_lock_bh()/__netif_tx_unlock_bh() instead.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agotcp: gso: do not generate out of order packets
Eric Dumazet [Wed, 15 May 2013 15:38:01 +0000 (15:38 +0000)]
tcp: gso: do not generate out of order packets

GSO TCP handler has following issues :

1) ooo_okay from original GSO packet is duplicated to all segments
2) segments (but the last one) are orphaned, so transmit path can not
get transmit queue number from the socket. This happens if GSO
segmentation is done before stacked device for example.

Result is we can send packets from a given TCP flow to different TX
queues (if using multiqueue NICS). This generates OOO problems and
spurious SACK & retransmits.

Fix this by keeping socket pointer set for all segments.

This means that every segment must also have a destructor, and the
original gso skb truesize must be split on all segments, to keep
precise sk->sk_wmem_alloc accounting.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Maciej Żenczykowski <maze@google.com>
Cc: Tom Herbert <therbert@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf
David S. Miller [Thu, 16 May 2013 21:32:42 +0000 (14:32 -0700)]
Merge branch 'master' of git://git./linux/kernel/git/pablo/nf

Pablo Neira Ayuso says:

====================
The following patchset contains three Netfilter fixes and update
for the MAINTAINER file for your net tree, they are:

* Fix crash if nf_log_packet is called from conntrack, in that case
  both interfaces are NULL, from Hans Schillstrom. This bug introduced
  with the logging netns support in the previous merge window.

* Fix compilation of nf_log and nf_queue without CONFIG_PROC_FS,
  from myself. This bug was introduced in the previous merge window
  with the new netns support for the netfilter logging infrastructure.

* Fix possible crash in xt_TCPOPTSTRIP due to missing sanity
  checkings to validate that the TCP header is well-formed, from
  myself. I can find this bug in 2.6.25, probably it's been there
  since the beginning. I'll pass this to -stable.

* Update MAINTAINER file to point to new nf trees at git.kernel.org,
  remove Harald and use M: instead of P: (now obsolete tag) to
  keep Jozsef in the list of people.

Please, consider pulling this. Thanks!
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoiwlwifi: mvm: fix aggregation drain flow
Emmanuel Grumbach [Tue, 7 May 2013 11:08:24 +0000 (14:08 +0300)]
iwlwifi: mvm: fix aggregation drain flow

Move the counter for non-AMPDU frames to mvm. It is needed
for the drain flow which happens once the ieee80211_sta has
been freed, so keeping it in iwl_mvm_sta which is embed into
ieee80211_sta is not a good idea.

Also, since its purpose it to remove the STA in the fw only
after all the frames for this station have exited the shared
Tx queues, we need to decrement it in the reclaim flow. This
flow can happen after ieee80211_sta has been removed, which
means that we have no iwl_mvm_sta there. So we can't know
what is the vif type. Hence, we know audit these frames for
all the vif types.
In order to avoid spawning sta_drained_wk all the time, we
now check that we are in a flow in which draining might
happen - only when mvmsta is NULL. This is better than
previous code that would spawn sta_drained_wk all the time
in AP mode.

Cc: stable@vger.kernel.org [3.9]
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Reviewed-by: Ilan Peer <ilan.peer@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
11 years agoiwlwifi: mvm: Prevent setting assoc flag in MAC_CONTEXT_CMD
Alexander Bondar [Thu, 2 May 2013 13:34:48 +0000 (16:34 +0300)]
iwlwifi: mvm: Prevent setting assoc flag in MAC_CONTEXT_CMD

In the normal flow first MAC_CONTEXT_CMD for particular interface is
never sent while associated. The exception is fw restart flow when
resuming from suspend when WoWLAN is enabled. In this case successive
"add" and "modify" MAC_CONTEXT_CMD commands may be sent with assoc flag
set what cause FW mal functioning. To prevent this never set assoc flag
in MAC_CONTEXT_CMD with action "add".

Cc: stable@vger.kernel.org
Signed-off-by: Alexander Bondar <alexander.bondar@intel.com>
Reviewed-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
11 years agoiwlwifi: mvm: tell firmware to let multicast frames in
Emmanuel Grumbach [Thu, 2 May 2013 12:01:24 +0000 (15:01 +0300)]
iwlwifi: mvm: tell firmware to let multicast frames in

Without this command, the firmware will filter out all the
multicast frames. Let them all in as for now. Later we will
want to optimize this to save power.

Cc: stable@vger.kernel.org
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
11 years agoiwlwifi: mvm: Always use SCAN_TYPE_FORCED
Ilan Peer [Sun, 28 Apr 2013 05:18:28 +0000 (08:18 +0300)]
iwlwifi: mvm: Always use SCAN_TYPE_FORCED

The FW AUX framework does not handle well cases where time events
fail to be scheduled (and as a result issues assert 0x3330). Until
a proper fix is in place, WA this by always setting the scan type to
SCAN_TYPE_FORCED.

Cc: stable@vger.kernel.org
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Reviewed-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
11 years agocfg80211: fix sending WoWLAN TCP wakeup settings
Johannes Berg [Thu, 16 May 2013 08:24:28 +0000 (10:24 +0200)]
cfg80211: fix sending WoWLAN TCP wakeup settings

The code sending the current WoWLAN TCP wakeup settings in
nl80211_send_wowlan_tcp() is not closing the nested attribute,
thus causing the parser to get confused on the receiver side
in userspace (iw). Fix this.

Cc: stable@vger.kernel.org [3.9]
Reported-by: Deepak Arora <deepakx.arora@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
11 years agomac80211: report deauth to cfg80211 for local state change
Johannes Berg [Tue, 14 May 2013 07:39:02 +0000 (09:39 +0200)]
mac80211: report deauth to cfg80211 for local state change

Even if the frame isn't transmitted to the AP, we need to
report it to cfg80211 so the state there can be updated
correctly.

Reported-by: Jouni Malinen <j@w1.fi>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
11 years agocfg80211: fix interface down/disconnect state handling
Johannes Berg [Tue, 14 May 2013 07:19:53 +0000 (09:19 +0200)]
cfg80211: fix interface down/disconnect state handling

When the interface goes down, there's no need to call
cfg80211_mlme_down() after __cfg80211_disconnect() as
the latter will call the former (if appropriate.)

Also, in __cfg80211_disconnect(), if the cfg80211 SME
isn't used, __cfg80211_disconnected() may still need
to be called (depending on the current state) so that
the SME state gets cleared.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
11 years agomac80211: fix AP-mode frame matching
Johannes Berg [Mon, 13 May 2013 14:42:40 +0000 (16:42 +0200)]
mac80211: fix AP-mode frame matching

In AP mode, ignore frames with mis-matched BSSID that aren't
multicast or sent to the correct destination. This fixes
reporting public action frames to userspace multiple times
on multiple virtual AP interfaces.

Cc: stable@vger.kernel.org
Reported-by: Jouni Malinen <j@w1.fi>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
11 years agonl80211: Add wdev identifier to some nl80211 notifications
Ilan Peer [Wed, 8 May 2013 13:35:55 +0000 (16:35 +0300)]
nl80211: Add wdev identifier to some nl80211 notifications

Adding the attributes fixes an issue with P2P Device not
working properly for management frame TX.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
11 years agomac80211: use just spin_lock() in ieee80211_get_tkip_p2k()
Johannes Berg [Mon, 6 May 2013 13:54:58 +0000 (15:54 +0200)]
mac80211: use just spin_lock() in ieee80211_get_tkip_p2k()

ieee80211_get_tkip_p2k() may be called with interrupts
disabled, so spin_unlock_bh() isn't safe and leads to
warnings. Since it's always called with BHs disabled
already, just use spin_lock().

Cc: stable@vger.kernel.org
Reported-by: Milan Kocian <milon@wq.cz>
Acked-by: Stanislaw Gruszka <sgruszka@redhat.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
11 years agomac80211: fix spurious RCU warning and update documentation
Felix Fietkau [Fri, 3 May 2013 08:01:03 +0000 (10:01 +0200)]
mac80211: fix spurious RCU warning and update documentation

Document rx vs tx status concurrency requirements.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
11 years agomac80211: fix HT beacon-based channel switch handling
Johannes Berg [Fri, 3 May 2013 07:35:35 +0000 (09:35 +0200)]
mac80211: fix HT beacon-based channel switch handling

When an HT AP is advertising channel switch in a beacon, it
doesn't (and shouldn't, according to 802.11-2012 Table 8-20)
include a secondary channel offset element. The only possible
interpretation is that the previous secondary channel offset
remains valid, so use that when switching channel based only
on beacon information.

VHT requires the Wide Bandwidth Channel Switch subelement to
be present in the Channel Switch Wrapper element, so the code
for that is probably ok (see 802.11ac Draft 4, 8.4.2.165.)

Reported-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
11 years agomac80211: fix IEEE80211_SDATA_DISCONNECT_RESUME
Johannes Berg [Mon, 29 Apr 2013 12:57:44 +0000 (14:57 +0200)]
mac80211: fix IEEE80211_SDATA_DISCONNECT_RESUME

Since commit 12e7f517029dad819c45eca9ca01fdb9ba57616b,
IEEE80211_SDATA_DISCONNECT_RESUME no longer worked as
it would simply never be tested. Restore a bit of the
code removed there and in 9b7d72c1041ec5b20b24af487a9
to make it work again.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
11 years agocfg80211: fix WoWLAN wakeup tracing
Johannes Berg [Mon, 29 Apr 2013 10:09:14 +0000 (12:09 +0200)]
cfg80211: fix WoWLAN wakeup tracing

If the device reports a non-wireless wakeup reason, the
tracing code crashes trying to dereference a NULL pointer.
Fix this by checking the pointer on all accesses and also
add a non_wireless tag to the event.

Cc: stable@vger.kernel.org
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
11 years agomac80211_hwsim: correctly register the platform driver
Sasha Levin [Wed, 24 Apr 2013 04:40:05 +0000 (00:40 -0400)]
mac80211_hwsim: correctly register the platform driver

Not registering a platform_driver would make us access garbage
when the platform callbacks under driver_register() kicks in.

Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
Tested-By: Martin Pitt <martin.pitt@ubuntu.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
11 years agocfg80211: fix wiphy_register error path
Johannes Berg [Fri, 10 May 2013 17:23:40 +0000 (19:23 +0200)]
cfg80211: fix wiphy_register error path

If rfkill_register() fails in wiphy_register() the struct device
is unregistered but everything else isn't (regulatory, debugfs)
and we even leave the wiphy instance on all internal lists even
though it will likely be freed soon, which is clearly a problem.
Fix this by cleaning up properly.

Cc: stable@vger.kernel.org
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
11 years agonetfilter: xt_TCPOPTSTRIP: fix possible mangling beyond packet boundary
Pablo Neira Ayuso [Tue, 7 May 2013 01:22:18 +0000 (03:22 +0200)]
netfilter: xt_TCPOPTSTRIP: fix possible mangling beyond packet boundary

This target assumes that tcph->doff is well-formed, that may be well
not the case. Add extra sanity checkings to avoid possible crash due
to read/write out of the real packet boundary. After this patch, the
default action on malformed TCP packets is to drop them. Moreover,
fragments are skipped.

Reported-by: Rafal Kupka <rkupka@telemetry.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
11 years agobroadcom: add include guards to include/linux/brcmphy.h
Florian Fainelli [Wed, 15 May 2013 08:00:25 +0000 (08:00 +0000)]
broadcom: add include guards to include/linux/brcmphy.h

include/linux/brcmphy.h is currently not protected against double
inclusion, add ifdefs guard to fix that.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agofec: Invert the order of function calls in fec_restart()
Fabio Estevam [Wed, 15 May 2013 07:06:27 +0000 (07:06 +0000)]
fec: Invert the order of function calls in fec_restart()

commit 54309fa6 ("net: fec: fix kernel oops when plug/unplug cable many times")
introduced the following 'if' block in the beginning of fec_start():

if (netif_running(ndev)) {
netif_device_detach(ndev);
napi_disable(&fep->napi);
netif_stop_queue(ndev);
netif_tx_lock_bh(ndev);
}

Then later in the end of fec_restart() there is another block that calls the
opposite of each one of these functions.

The correct approach would be to also call them with in the reverse order, so
that we have as result:

if (netif_running(ndev)) {
netif_tx_unlock_bh(ndev);
netif_wake_queue(ndev);
napi_enable(&fep->napi);
netif_device_attach(ndev);
}

Suggested-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agofec: Fix inconsistent lock state
Fabio Estevam [Wed, 15 May 2013 07:06:26 +0000 (07:06 +0000)]
fec: Fix inconsistent lock state

fec_restart() runs in softirq context and we should use the
netif_tx_lock_bh/netif_tx_unlock_bh() variants to avoid the following warning
that happens since commit 54309fa6 ("net: fec: fix kernel oops when plug/unplug
cable many times"):

[    9.753168] =================================
[    9.757540] [ INFO: inconsistent lock state ]
[    9.761921] 3.10.0-rc1-next-20130514 #13 Not tainted
[    9.766897] ---------------------------------
[    9.771264] inconsistent {SOFTIRQ-ON-W} -> {IN-SOFTIRQ-W} usage.
[    9.777288] swapper/0 [HC0[0]:SC1[3]:HE1:SE0] takes:
[    9.782261]  (_xmit_ETHER#2){+.?...}, at: [<c03c24a4>] sch_direct_xmit+0xa0/0x2d4
[    9.789879] {SOFTIRQ-ON-W} state was registered at:
[    9.794769]   [<c0059c60>] __lock_acquire+0x528/0x1bc0
[    9.799953]   [<c005b838>] lock_acquire+0xa0/0x108
[    9.804780]   [<c0441320>] _raw_spin_lock+0x28/0x38
[    9.809702]   [<c02f9fc8>] fec_restart+0x5d0/0x664
[    9.814542]   [<c02fa738>] fec_enet_adjust_link+0xa8/0xc0
[    9.819978]   [<c02f7a28>] phy_state_machine+0x2fc/0x370
[    9.825323]   [<c0035ee0>] process_one_work+0x1c0/0x4a0
[    9.830589]   [<c0036594>] worker_thread+0x138/0x394
[    9.835587]   [<c003c620>] kthread+0xa4/0xb0
[    9.839890]   [<c000e820>] ret_from_fork+0x14/0x34
[    9.844728] irq event stamp: 185984
[    9.848226] hardirqs last  enabled at (185984): [<c00232b0>] local_bh_enable_ip+0x84/0xf0
[    9.856450] hardirqs last disabled at (185983): [<c0023270>] local_bh_enable_ip+0x44/0xf0
[    9.864667] softirqs last  enabled at (185966): [<c0023470>] irq_enter+0x64/0x68
[    9.872099] softirqs last disabled at (185967): [<c0023510>] irq_exit+0x9c/0xd8
[    9.879440]
[    9.879440] other info that might help us debug this:
[    9.885981]  Possible unsafe locking scenario:
[    9.885981]
[    9.891912]        CPU0
[    9.894364]        ----
[    9.896814]   lock(_xmit_ETHER#2);
[    9.900259]   <Interrupt>
[    9.902884]     lock(_xmit_ETHER#2);
[    9.906500]
[    9.906500]  *** DEADLOCK ***

Reported-by: Shawn Guo <shawn.guo@linaro.org>
Suggested-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoipg: fix an unsigned widening cast of '~' truncation issue
Dan Carpenter [Tue, 14 May 2013 21:20:02 +0000 (21:20 +0000)]
ipg: fix an unsigned widening cast of '~' truncation issue

The bug here is this code from ipg_nic_hard_start_xmit():

txfd->tfc &= cpu_to_le64(~IPG_TFC_TFDDONE);

IPG_TFC_TFDDONE is 0x0000000080000000 so it's an unsigned int.  The
negated value is 0x7fffffff but 0xffffffff7fffffff was intended.

The other values in this file don't need to be changed but I did it for
consistency.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonetfilter: log: netns NULL ptr bug when calling from conntrack
Hans Schillstrom [Wed, 15 May 2013 01:23:45 +0000 (01:23 +0000)]
netfilter: log: netns NULL ptr bug when calling from conntrack

Since (69b34fb netfilter: xt_LOG: add net namespace support
for xt_LOG), we hit this:

[ 4224.708977] BUG: unable to handle kernel NULL pointer dereference at 0000000000000388
[ 4224.709074] IP: [<ffffffff8147f699>] ipt_log_packet+0x29/0x270

when callling log functions from conntrack both in and out
are NULL i.e. the net pointer is invalid.

Adding struct net *net in call to nf_logfn() will secure that
there always is a vaild net ptr.

Reported as netfilter's bugzilla bug 818:
https://bugzilla.netfilter.org/show_bug.cgi?id=818

Reported-by: Ronald <ronald645@gmail.com>
Signed-off-by: Hans Schillstrom <hans@schillstrom.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
11 years agonet/macb: fix ISR clear-on-write behavior only for some SoC
Nicolas Ferre [Tue, 14 May 2013 03:00:16 +0000 (03:00 +0000)]
net/macb: fix ISR clear-on-write behavior only for some SoC

Commit 749a2b6 (net/macb: clear tx/rx completion flags in ISR)
introduces clear-on-write on ISR register. This behavior is not always
implemented when using Cadence MACB/GEM and is breaking other platforms.
We are using the Design Configuration Register 1 information and a capability
property to actually activate this clear-on-write behavior on ISR.

Reported-by: Hein Tibosch <hein_tibosch@yahoo.es>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Tested-by: Hein Tibosch <hein_tibosch@yahoo.es>
Acked-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agonet/802/mrp: fix lockdep splat
Eric Dumazet [Mon, 13 May 2013 02:24:11 +0000 (02:24 +0000)]
net/802/mrp: fix lockdep splat

commit fb745e9a037895 ("net/802/mrp: fix possible race condition when
calling mrp_pdu_queue()") introduced a lockdep splat.

[   19.735147] =================================
[   19.735235] [ INFO: inconsistent lock state ]
[   19.735324] 3.9.2-build-0063 #4 Not tainted
[   19.735412] ---------------------------------
[   19.735500] inconsistent {IN-SOFTIRQ-W} -> {SOFTIRQ-ON-W} usage.
[   19.735592] rmmod/1840 [HC0[0]:SC0[0]:HE1:SE1] takes:
[   19.735682]  (&(&app->lock)->rlock#2){+.?...}, at: [<f862bb5b>]
mrp_uninit_applicant+0x69/0xba [mrp]

app->lock is normally taken under softirq context, so disable BH to
avoid the splat.

Reported-by: Denys Fedoryshchenko <denys@visp.net.lb>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: David Ward <david.ward@ll.mit.edu>
Cc: Cong Wang <amwang@redhat.com>
Tested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agomv643xx_eth: fix NAPI weight being > 64
Andrew Lunn [Mon, 13 May 2013 23:54:20 +0000 (23:54 +0000)]
mv643xx_eth: fix NAPI weight being > 64

3.10-rc1 issues the following warning:

netif_napi_add() called with weight 128 on device eth%d

This patch reduce the weight to 64, using NAPI_POLL_WEIGHT.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agotcp: fix tcp_md5_hash_skb_data()
Eric Dumazet [Mon, 13 May 2013 21:25:52 +0000 (21:25 +0000)]
tcp: fix tcp_md5_hash_skb_data()

TCP md5 communications fail [1] for some devices, because sg/crypto code
assume page offsets are below PAGE_SIZE.

This was discovered using mlx4 driver [2], but I suspect loopback
might trigger the same bug now we use order-3 pages in tcp_sendmsg()

[1] Failure is giving following messages.

huh, entered softirq 3 NET_RX ffffffff806ad230 preempt_count 00000100,
exited with 00000101?

[2] mlx4 driver uses order-2 pages to allocate RX frags

Reported-by: Matt Schnall <mischnal@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Bernhard Beck <bbeck@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agobe2net: Avoid double insertion of vlan tags.
Sarveshwar Bandi [Mon, 13 May 2013 20:28:20 +0000 (20:28 +0000)]
be2net: Avoid double insertion of vlan tags.

Fix to avoid double insertion of  vlan tags into the packet while
handling an asic workaroud (issue introduced by net next Commit bc0c340)

Signed-off-by: Sarveshwar Bandi <sarveshwar.bandi@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agosfc: Reduce RX scatter buffer size, and reduce alignment if appropriate
Ben Hutchings [Mon, 13 May 2013 12:01:22 +0000 (12:01 +0000)]
sfc: Reduce RX scatter buffer size, and reduce alignment if appropriate

efx_start_datapath() asserts that we can fit 2 RX scatter buffers plus
a software structure, each appropriately aligned, into a single page.
Where L1_CACHE_BYTES == 256 and PAGE_SIZE == 4096, which is the case
on s390, this assertion fails.

The current scatter buffer size is also not a multiple of 64 or 128,
which are more common cache line sizes.  If we can make both the start
and end of a scatter buffer cache-aligned, this will reduce the need
for read-modify-write operations on inter- processor links.

Fix the alignment by reducing EFX_RX_USR_BUF_SIZE to 2048 - 256 ==
1792.  (We could use 2048 - L1_CACHE_BYTES, but EFX_RX_USR_BUF_SIZE
also affects user-level networking where a larger amount of
housekeeping data may be needed.  Although this version of the driver
does not support user-level networking, I prefer to keep scattering
behaviour consistent with the out-of-tree version.)

This still doesn't fix the s390 build because like most architectures
it has NET_IP_ALIGN == 2.  When NET_IP_ALIGN != 0 we cannot achieve
cache line alignment at either the start or end of a scatter buffer,
so there is actually no point in padding the buffers to a multiple of
the cache line size.  All we need is 4-byte alignment of the network
header, so do that.

Adjust the assertions accordingly.

Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
Reported-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Acked-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agosfc: Delete EFX_PAGE_IP_ALIGN, equivalent to NET_IP_ALIGN
Ben Hutchings [Mon, 13 May 2013 11:58:31 +0000 (11:58 +0000)]
sfc: Delete EFX_PAGE_IP_ALIGN, equivalent to NET_IP_ALIGN

The two architectures that define CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS
(powerpc and x86) now both define NET_IP_ALIGN as 0, so there is no
need for this optimisation any more.

Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agotg3: Fix data corruption on 5725 with TSO
Michael Chan [Mon, 13 May 2013 11:04:16 +0000 (11:04 +0000)]
tg3: Fix data corruption on 5725 with TSO

The 5725 family of devices (asic rev 5762), corrupts TSO packets where
the buffer is within MSS bytes of a 4G boundary (4G, 8G etc.). Detect
this condition and trigger the workaround path.

Cc: <stable@vger.kernel.org>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agotg3: Skip powering down function 0 on certain serdes devices
Nithin Sujir [Mon, 13 May 2013 11:04:15 +0000 (11:04 +0000)]
tg3: Skip powering down function 0 on certain serdes devices

On the 5718, 5719 and 5720 serdes devices, powering down function 0
results in all the other ports being powered down. Add code to skip
function 0 power down.

v2:
 - Modify tg3_phy_power_bug() function to use a switch instead of a
   complicated if statement. Suggested by Joe Perches.

Cc: <stable@vger.kernel.org>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Mon, 13 May 2013 20:25:36 +0000 (13:25 -0700)]
Merge git://git./linux/kernel/git/davem/net

Pull networking fixes from David Miller:
 "Several small bug fixes all over:

   1) be2net driver uses wrong payload length when submitting MAC list
      get requests to the chip.  From Sathya Perla.

   2) Fix mwifiex memory leak on driver unload, from Amitkumar Karwar.

   3) Prevent random memory access in batman-adv, from Marek Lindner.

   4) batman-adv doesn't check for pskb_trim_rcsum() errors, also from
      Marek Lindner.

   5) Fix fec crashes on rapid link up/down, from Frank Li.

   6) Fix inner protocol grovelling in GSO, from Pravin B Shelar.

   7) Link event validation fix in qlcnic from Rajesh Borundia.

   8) Not all FEC chips can support checksum offload, fix from Shawn
      Guo.

   9) EXPORT_SYMBOL + inline doesn't make any sense, from Denis Efremov.

  10) Fix race in passthru mode during device removal in macvlan, from
      Jiri Pirko.

  11) Fix RCU hash table lookup socket state race in ipv6, leading to
      NULL pointer derefs, from Eric Dumazet.

  12) Add several missing HAS_DMA kconfig dependencies, from Geert
      Uyttterhoeven.

  13) Fix bogus PCI resource management in 3c59x driver, from Sergei
      Shtylyov.

  14) Fix info leak in ipv6 GRE tunnel driver, from Amerigo Wang.

  15) Fix device leak in ipv6 IPSEC policy layer, from Cong Wang.

  16) DMA mapping leak fix in qlge from Thadeu Lima de Souza Cascardo.

  17) Missing iounmap on probe failure in bna driver, from Wei Yongjun."

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (40 commits)
  bna: add missing iounmap() on error in bnad_init()
  qlge: fix dma map leak when the last chunk is not allocated
  xfrm6: release dev before returning error
  ipv6,gre: do not leak info to user-space
  virtio_net: use default napi weight by default
  emac: Fix EMAC soft reset on 460EX/GT
  3c59x: fix PCI resource management
  caif: CAIF_VIRTIO should depend on HAS_DMA
  net/ethernet: MACB should depend on HAS_DMA
  net/ethernet: ARM_AT91_ETHER should depend on HAS_DMA
  net/wireless: ATH9K should depend on HAS_DMA
  net/ethernet: STMMAC_ETH should depend on HAS_DMA
  net/ethernet: NET_CALXEDA_XGMAC should depend on HAS_DMA
  ipv6: do not clear pinet6 field
  macvlan: fix passthru mode race between dev removal and rx path
  ipv4: ip_output: remove inline marking of EXPORT_SYMBOL functions
  net/mlx4: Strengthen VLAN tags/priorities enforcement in VST mode
  net/mlx4_core: Add missing report on VST and spoof-checking dev caps
  net: fec: enable hardware checksum only on imx6q-fec
  qlcnic: Fix validation of link event command.
  ...

11 years agobna: add missing iounmap() on error in bnad_init()
Wei Yongjun [Mon, 13 May 2013 04:26:06 +0000 (04:26 +0000)]
bna: add missing iounmap() on error in bnad_init()

Add the missing iounmap() before return from bnad_init()
in the error handling case.
Introduced by commit 01b54b1451853593739816a392485c4e2bee7dda
(bna: tx rx cleanup fix).

Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoqlge: fix dma map leak when the last chunk is not allocated
Thadeu Lima de Souza Cascardo [Sat, 11 May 2013 09:15:37 +0000 (09:15 +0000)]
qlge: fix dma map leak when the last chunk is not allocated

qlge allocates chunks from a page that it maps and unmaps that page when
the last chunk is released. When the driver is unloaded or the card is
removed, all chunks are released and the page is unmapped for the last
chunk.

However, when the last chunk of a page is not allocated and the device
is removed, that page is not unmapped. In fact, its last reference is
not put and there's also a page leak. This bug prevents a device from
being properly hotplugged.

When the DMA API debug option is enabled, the following messages show
the pending DMA allocation after we remove the driver.

This patch fixes the bug by unmapping and putting the page from the ring
if its last chunk has not been allocated.

pci 0005:98:00.0: DMA-API: device driver has pending DMA allocations while released from device [count=1]
One of leaked entries details: [device address=0x0000000060a80000] [size=65536 bytes] [mapped with DMA_FROM_DEVICE] [mapped as page]
------------[ cut here ]------------
WARNING: at lib/dma-debug.c:746
Modules linked in: qlge(-) rpadlpar_io rpaphp pci_hotplug fuse [last unloaded: qlge]
NIP: c0000000003fc3ec LR: c0000000003fc3e8 CTR: c00000000054de60
REGS: c0000003ee9c74e0 TRAP: 0700   Tainted: G           O  (3.7.2)
MSR: 8000000000029032 <SF,EE,ME,IR,DR,RI>  CR: 28002424  XER: 00000001
SOFTE: 1
CFAR: c0000000007a39c8
TASK = c0000003ee8d5c90[8406] 'rmmod' THREAD: c0000003ee9c4000 CPU: 31
GPR00: c0000000003fc3e8 c0000003ee9c7760 c000000000c789f8 00000000000000ee
GPR04: 0000000000000000 00000000000000ef 0000000000004000 0000000000010000
GPR08: 00000000000000be c000000000b22088 c000000000c4c218 00000000007c0000
GPR12: 0000000028002422 c00000000ff26c80 0000000000000000 000001001b0f1b40
GPR16: 00000000100cb9d8 0000000010093088 c000000000cdf910 0000000000000001
GPR20: 0000000000000000 c000000000dbfc00 0000000000000000 c000000000dbfb80
GPR24: c0000003fafc9d80 0000000000000001 000000000001ff80 c0000003f38f7888
GPR28: c000000000ddfc00 0000000000000400 c000000000bd7790 c000000000ddfb80
NIP [c0000000003fc3ec] .dma_debug_device_change+0x22c/0x2b0
LR [c0000000003fc3e8] .dma_debug_device_change+0x228/0x2b0
Call Trace:
[c0000003ee9c7760] [c0000000003fc3e8] .dma_debug_device_change+0x228/0x2b0 (unreliable)
[c0000003ee9c7840] [c00000000079a098] .notifier_call_chain+0x78/0xf0
[c0000003ee9c78e0] [c0000000000acc20] .__blocking_notifier_call_chain+0x70/0xb0
[c0000003ee9c7990] [c0000000004a9580] .__device_release_driver+0x100/0x140
[c0000003ee9c7a20] [c0000000004a9708] .driver_detach+0x148/0x150
[c0000003ee9c7ac0] [c0000000004a8144] .bus_remove_driver+0xc4/0x150
[c0000003ee9c7b60] [c0000000004aa58c] .driver_unregister+0x8c/0xe0
[c0000003ee9c7bf0] [c0000000004090b4] .pci_unregister_driver+0x34/0xf0
[c0000003ee9c7ca0] [d000000002231194] .qlge_exit+0x1c/0x34 [qlge]
[c0000003ee9c7d20] [c0000000000e36d8] .SyS_delete_module+0x1e8/0x290
[c0000003ee9c7e30] [c0000000000098d4] syscall_exit+0x0/0x94
Instruction dump:
7f26cb78 e818003a e87e81a0 e8f80028 e9180030 796b1f24 78001f24 7d6a5a14
7d2a002a e94b0020 483a7595 60000000 <0fe000002fb80000 40de0048 80120050
---[ end trace 4294f9abdb01031d ]---
Mapped at:
 [<d000000002222f54>] .ql_update_lbq+0x384/0x580 [qlge]
 [<d000000002227bd0>] .ql_clean_inbound_rx_ring+0x300/0xc60 [qlge]
 [<d0000000022288cc>] .ql_napi_poll_msix+0x39c/0x5a0 [qlge]
 [<c0000000006b3c50>] .net_rx_action+0x170/0x300
 [<c000000000081840>] .__do_softirq+0x170/0x300

Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@linux.vnet.ibm.com>
Acked-by: Jitendra Kalsaria <Jitendra.kalsaria@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoMerge tag 'spi-v3.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi
Linus Torvalds [Mon, 13 May 2013 15:12:18 +0000 (08:12 -0700)]
Merge tag 'spi-v3.10-rc1' of git://git./linux/kernel/git/broonie/spi

Pull spi updates from Mark Brown:
 "A few driver specific fixes plus improved error handling in the
  generic DT GPIO chipselect handling - not exciting but useful."

* tag 'spi-v3.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi:
  spi/spi-atmel: BUG: fix doesn' support 16 bits transfers using PIO
  spi/davinci: fix module build error
  spi: Return error from of_spi_register_master on bad "cs-gpios" property
  spi: Initialize cs_gpio and cs_gpios with -ENOENT
  spi/atmel: fix speed_hz check in atmel_spi_transfer()

11 years agoMerge branch 'drm-next' of git://people.freedesktop.org/~airlied/linux
Linus Torvalds [Mon, 13 May 2013 14:59:59 +0000 (07:59 -0700)]
Merge branch 'drm-next' of git://people.freedesktop.org/~airlied/linux

Pull drm fixes from Dave Airlie:
 "Just a few straggling fixes I hoovered up, and an intel fixes pull
  from Daniel which fixes some regressions, and some mgag200 fixes from
  Matrox."

* 'drm-next' of git://people.freedesktop.org/~airlied/linux:
  drm/mgag200: Fix framebuffer base address programming
  drm/mgag200: Convert counter delays to jiffies
  drm/mgag200: Fix writes into MGA1064_PIX_CLK_CTL register
  drm/mgag200: Don't change unrelated registers during modeset
  drm: Only print a debug message when the polled connector has changed
  drm: Make the HPD status updates debug logs more readable
  drm: Use names of ioctls in debug traces
  drm: Remove pointless '-' characters from drm_fb_helper documentation
  drm: Add kernel-doc for drm_fb_helper_funcs->initial_config
  drm: refactor call to request_module
  drm: Don't prune modes loudly when a connector is disconnected
  drm: Add missing break in the command line mode parsing code
  drm/i915: clear the stolen fb before resuming
  Revert "drm/i915: Calculate correct stolen size for GEN7+"
  drm/i915: hsw: fix link training for eDP on port-A
  Revert "drm/i915: revert eDP bpp clamping code changes"
  drm: don't check modeset locks in panic handler
  drm/i915: Fix pipe enabled mask for pipe C in WM calculations
  drm/mm: fix dump table BUG
  drm/i915: Always normalize return timeout for wait_timeout_ioctl

11 years agoMerge tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rusty...
Linus Torvalds [Mon, 13 May 2013 14:59:08 +0000 (07:59 -0700)]
Merge tag 'fixes-for-linus' of git://git./linux/kernel/git/rusty/linux

Pull virtio/lguest fixes from Rusty Russell:
 "Missing license tag and some fallout from the lguest pagetable rework"

* tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rusty/linux:
  lguest: clear cached last cpu when guest_set_pgd() called.
  Add missing module license tag to vring helpers.

11 years agoMerge remote-tracking branch 'spi/fix/grant' into spi-linus
Mark Brown [Mon, 13 May 2013 14:27:18 +0000 (18:27 +0400)]
Merge remote-tracking branch 'spi/fix/grant' into spi-linus

11 years agoMerge remote-tracking branch 'spi/fix/atmel' into spi-linus
Mark Brown [Mon, 13 May 2013 14:27:16 +0000 (18:27 +0400)]
Merge remote-tracking branch 'spi/fix/atmel' into spi-linus

11 years agodrm/mgag200: Fix framebuffer base address programming
Christopher Harvey [Wed, 8 May 2013 19:10:38 +0000 (19:10 +0000)]
drm/mgag200: Fix framebuffer base address programming

Higher bits of the base address of framebuffers weren't being
programmed properly. This caused framebuffers that didn't happen to be
allocated at a low enough address to not be displayed properly.

Signed-off-by: Christopher Harvey <charvey@matrox.com>
Signed-off-by: Mathieu Larouche <mathieu.larouche@matrox.com>
Acked-by: Julia Lemire <jlemire@matrox.com>
Tested-by: Julia Lemire <jlemire@matrox.com>
Cc: stable@vger.kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>
11 years agodrm/mgag200: Convert counter delays to jiffies
Christopher Harvey [Mon, 6 May 2013 15:56:17 +0000 (15:56 +0000)]
drm/mgag200: Convert counter delays to jiffies

Signed-off-by: Christopher Harvey <charvey@matrox.com>
Acked-by: Julia Lemire <jlemire@matrox.com>
Tested-by: Julia Lemire <jlemire@matrox.com>
Acked-by: Mathieu Larouche <mathieu.larouche@matrox.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
11 years agodrm/mgag200: Fix writes into MGA1064_PIX_CLK_CTL register
Christopher Harvey [Fri, 12 Apr 2013 22:24:05 +0000 (22:24 +0000)]
drm/mgag200: Fix writes into MGA1064_PIX_CLK_CTL register

The original line,
  WREG_DAC(MGA1064_PIX_CLK_CTL_CLK_DIS, tmp);
wrote tmp into MGA1064_PIX_CLK_CTL_CLK_DIS, where
MGA1064_PIX_CLK_CTL_CLK_DIS is an offset into
MGA1064_PIX_CLK_CTL. Change the line to write properly into
MGA1064_PIX_CLK_CTL. There were other chunks of code nearby that use
the same pattern (but work correctly), so this patch updates them all
to use this new (slightly more efficient) write pattern. The WREG_DAC
macro was causing the DAC_INDEX register to be set to the same value
twice. WREG8(DAC_DATA, foo) takes advantage of the fact that DAC_INDEX
is already at the value we want.

Signed-off-by: Christopher Harvey <charvey@matrox.com>
Acked-by: Julia Lemire <jlemire@matrox.com>
Tested-by: Julia Lemire <jlemire@matrox.com>
Acked-by: Mathieu Larouche <mathieu.larouche@matrox.com>
Cc: stable@vger.kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>