drivers: hv: Cleanup the kvp related state in hyperv.h

Now cleanup the hyperv.h with regards to KVP definitions.

Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
(cherry picked from commit 59a084a70afa0678bda2a23a7bc7cc59664945c7)

Signed-off-by: Andy Whitcroft <apw@canonical.com>

tools: hv: Use hyperv.h to get the KVP definitions

Now use hyperv.h to get the KVP defines in the KVP user-mode code.

Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
(cherry picked from commit eab6af71f0b83a7f62b9c48be5b2c0a82a86fad3)

Signed-off-by: Andy Whitcroft <apw@canonical.com>

net/hyperv: Remove the unnecessary memset in rndis_filter_send()

The memory has been allocated by kzalloc, so it's unnecessary to memset
again.

Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
(cherry picked from commit bce60806de882cf15ad3a80a51e9878863a8fced)

Signed-off-by: Andy Whitcroft <apw@canonical.com>

net/hyperv: Correct the assignment in netvsc_recv_callback()

The first assignment to variable "net" is wrong, but overridden by the
latter assignments. So the bug isn't manifested.

Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
(cherry picked from commit 6f4c44460750dd4eb9926a58ab1ad0ceacef8284)

Signed-off-by: Andy Whitcroft <apw@canonical.com>

net/hyperv: Convert camel cased variables in rndis_filter.c to lower cases

Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Reviewed-by: Jesper Juhl <jj@chaosbits.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
(cherry picked from commit 5fccab3b66d53883a97fc65e0c33f3ebf74e8ff9)

Signed-off-by: Andy Whitcroft <apw@canonical.com>

drivers: hv: kvp: Move the contents of hv_kvp.h to hyperv.h

In preparation for consolidating all KVP related defines into a single header file
that both the kernel and user level components can use, move the contents of
hv_kvp.h into hyperv.h.

Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
(cherry picked from commit 2939437ce8f2de07237eb2bcce29b6a699bfe799)

Signed-off-by: Andy Whitcroft <apw@canonical.com>

drivers: hv: kvp: Add/cleanup connector defines

The current KVP code carries some private connector related defines.
Update connector.h to have all the KVP defines. As part of this patch
get rid of some unused defines.

Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
(cherry picked from commit 4f03a2c934894f30a64d397df8c7c4de129c5b30)

Signed-off-by: Andy Whitcroft <apw@canonical.com>

drivers: hv: Get rid of some unnecessary code

The current code unnecessarily limits the number of offers we handle.
Get rid of this limitation. As part of this cleanup, also get rid of an
unused define - MAX_MSG_TYPES.

Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
(cherry picked from commit 2897a563a55442379e5e59ec68c229a7f27fb7c6)

Signed-off-by: Andy Whitcroft <apw@canonical.com>

drm/i915: Sanitize BIOS debugging bits from PIPECONF

Quoting the BSpec from time immemorial:

  PIPEACONF, bits 28:27: Frame Start Delay (Debug)

  Used to delay the frame start signal that is sent to the display planes.
  Care must be taken to insure that there are enough lines during VBLANK
  to support this setting.

An instance of the BIOS leaving these bits set was found in the wild,
where it caused our modesetting to go all squiffy and skewiff.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=47271
Reported-and-tested-by: Eva Wang <evawang@linpus.com>
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=43012
Reported-and-tested-by: Carl Richell <carl@system76.com>
Cc: stable@kernel.org
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

(cherry-picked from commit f47166d2b0001fcb752b40c5a2d4db986dfbea68 drm-intel)
BugLink: http://bugs.launchpad.net/bugs/968307
Acked-by: Robert Hooker <sarvatt@canonical.com>
Signed-off-by: Andy Whitcroft <apw@canonical.com>

UBUNTU: SAUCE: (drop after 3.5) drm/i915: reinstate GM45 TV detection fix

BugLink: http://bugs.launchpad.net/bugs/638939
cherry picked from drm-intel-next-queue

This reverts commmit d4b74bf07873da2e94219a7b67a334fc1c3ce649 which
reverted the origin fix fb8b5a39b6310379d7b54c0c7113703a8eaf4a57.

We have at least 3 different bug reports that this fixes things and no
indication what is exactly wrong with this. So try again.

To make matters slightly more fun, the commit itself was cc: stable
whereas the revert has not been.

According to Peter Clifton he discussed this with Zhao Yakui and this
seems to be in contradiction of the GM45 PRM, but rumours have it that
this is how the BIOS does it ... let's see.

Reviewed-by: Rodrigo Vivi <rodrigo.vivi@gmail.com>
Tested-by: Peter Clifton <Peter.Clifton@clifton-electronics.com>
Cc: Zhao Yakui <yakui.zhao@intel.com>
Cc: Dave Airlie <airlied@redhat.com>
Cc: Eric Anholt <eric@anholt.net>
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=16236
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=25913
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=14792
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Brad Figg <brad.figg@canonical.com>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

UBUNTU: [Config] Enable CONFIG_USBIP_CORE=m

BugLink: http://bugs.launchpad.net/bugs/900384

CONFIG_USBIP_CORE=m
CONFIG_USBIP_HOST=m
CONFIG_USBIP_VHCI_HCD=m

Signed-off-by: Leann Ogasawara <leann.ogasawara@canonical.com>

eeepc-wmi: refine quirks handling

BugLink: http://bugs.launchpad.net/bugs/969576

Signed-off-by: Corentin Chary <corentin.chary@gmail.com>
Signed-off-by: Matthew Garrett <mjg@redhat.com>
(cherry picked from commit c55d995dd3cebffdeb2b7eff8acc813c56d62c97)

Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

asus-wmi: move WAPF variable into quirks_entry

BugLink: http://bugs.launchpad.net/bugs/969576

Some models work better with different values of wapf, so move the
variable into quriks_entry to make it more easy to give a specific
value to different models.

Based on original patch from AceLan Kao <acelan.kao@canonical.com>

Cc: AceLan Kao <acelan.kao@canonical.com>
Signed-off-by: Corentin Chary <corentin.chary@gmail.com>
Signed-off-by: Matthew Garrett <mjg@redhat.com>
(cherry picked from commit 6a2bcccdb3206950d28e343476d9050e23e79b7e)

Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

ACPICA: Fix regression in FADT revision checks

BugLink: http://bugs.launchpad.net/bugs/883441

commit 64b3db22c04586997ab4be46dd5a5b99f8a2d390 (2.6.39),
"Remove use of unreliable FADT revision field" causes regression
for old P4 systems because now cst_control and other fields are
not reset to 0.

The effect is that acpi_processor_power_init will notice
cst_control != 0 and a write to CST_CNT register is performed
that should not happen. As result, the system oopses after the
"No _CST, giving up" message, sometimes in acpi_ns_internalize_name,
sometimes in acpi_ns_get_type, usually at random places. May be
during migration to CPU 1 in acpi_processor_get_throttling.

Every one of these settings help to avoid this problem:
 - acpi=off
 - processor.nocst=1
 - maxcpus=1

The fix is to update acpi_gbl_FADT.header.length after
the original value is used to check for old revisions.

https://bugzilla.kernel.org/show_bug.cgi?id=42700
https://bugzilla.redhat.com/show_bug.cgi?id=727865

Signed-off-by: Julian Anastasov <ja@ssi.bg>
Acked-by: Bob Moore <robert.moore@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
(cherry picked from commit 3e80acd1af40fcd91a200b0416a7616b20c5d647)

Signed-off-by: Colin King <colin.king@canonical.com>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

UBUNTU: SAUCE: SECCOMP: adjust prctl constant

3.4-rc1 contains new prctl values that conflict with the ones chosen for
nnp, so push them forward with the hopes that we will align with the future.

Signed-off-by: Kees Cook <kees@ubuntu.com>
Signed-off-by: Leann Ogasawara <leann.ogasawara@canonical.com>

UBUNTU: SAUCE: SECCOMP: audit: always report seccomp violations

Violations of seccomp filters should always be reported, regardless
of audit context. This the minimal change version of what has been
proposed upstream: https://lkml.org/lkml/2012/3/23/332

Signed-off-by: Kees Cook <kees@ubuntu.com>
Signed-off-by: Leann Ogasawara <leann.ogasawara@canonical.com>

Revert "sched: tg->se->load should be initialised to tg->shares"

Dropping since this patch never made it upstream.

This reverts commit f55c2f6c615f6c975eae2d956b54bbc3e81de2f6.

sched: Fix nohz load accounting

BugLink: http://bugs.launchpad.net/bugs/838811

Various people reported nohz load tracking still being wrecked, but Doug
spotted the actual problem. We fold the nohz remainder in too soon,
causing us to loose samples and under-account.

So instead of playing catch-up up-front, always do a single load-fold
with whatever state we encounter and only then fold the nohz remainder
and play catch-up.

Reported-by: Doug Smythies <dsmythies@telus.net>
Reported-by: LesÅ=82aw Kope=C4=87 <leslaw.kopec@nasza-klasa.pl>
Reported-by: Aman Gupta <aman@tmm1.net>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link:
http://lkml.kernel.org/n/tip-4v31etnhgg9kwd6ocgx3rxl8@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@elte.hu>

(backport of upstream: c308b56b5398779cd3da0f62ab26b0453494c3d4)

Signed-off-by: Brad Figg <brad.figg@canonical.com>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

ethtool: Null-terminate filename passed to ethtool_ops::flash_device

BugLink: http://bugs.launchpad.net/bugs/921793

The parameters for ETHTOOL_FLASHDEV include a filename, which ought to
be null-terminated.  Currently the only driver that implements
ethtool_ops::flash_device attempts to add a null terminator if
necessary, but does it wrongly.  Do it in the ethtool core instead.

Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: create RSS rings even in multi-channel configs

BugLink: http://bugs.launchpad.net/bugs/921793

Currently RSS rings are not created in a multi-channel config.
RSS rings can be created on one (out of four) interfaces per port in a
multi-channel config. Doing this insulates the driver from a FW bug wherin
multi-channel config is wrongly reported even when not enabled. This also
helps performance in a multi-channel config, as one interface per port gets
RSS rings.

Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

netdev: make net_device_ops const

BugLink: http://bugs.launchpad.net/bugs/921793

More drivers where net_device_ops should be const.

Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: query link status in be_open()

BugLink: http://bugs.launchpad.net/bugs/921793

be2net gets an async link status notification from the FW when it creates
an MCC queue. There are some cases in which this gratuitous notification
is not received from FW. To cover this explicitly query the link status
in be_open().

Signed-off-by: Vasundhara Volam <vasundhara.volam@emulex.com>
Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
Signed-off-by: Ajit Khaparde <ajit.khaparde@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: fix range check for set_qos for a VF

BugLink: http://bugs.launchpad.net/bugs/921793

Signed-off-by: Ajit Khaparde <ajit.khaparde@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: Fix INTx processing for Lancer

BugLink: http://bugs.launchpad.net/bugs/921793

Lancer does not have HW registers to indicate the EQ causing the INTx
interrupt. As a result EQE entries of one EQ may be consumed when interrupt
is caused by another EQ. Fix this by arming CQs at the end of NAPI poll
routine to regenerate the EQEs.

Signed-off-by: Padmanabh Ratnakar <padmanabh.ratnakar@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: Add support for Skyhawk cards

BugLink: http://bugs.launchpad.net/bugs/921793

Signed-off-by: Ajit Khaparde <ajit.khaparde@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: refactor/cleanup vf configuration code

BugLink: http://bugs.launchpad.net/bugs/921793

- use adapter->num_vfs (and not the module param) to store the actual
number of vfs created. Use the same variable to reflect SRIOV
enable/disable state. So, drop the adapter->sriov_enabled field.

- use for_all_vfs() macro in VF configuration code

- drop the "vf_" prefix for the fields of be_vf_cfg; the prefix is
redundant and removing it helps reduce line wrap

Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: fix ethtool ringparam reporting

BugLink: http://bugs.launchpad.net/bugs/921793

The ethtool "-g" option is supposed to report the max queue length and
user modified queue length for RX and TX queues.  be2net doesn't support
user modification of queue lengths. So, the correct values for these
would be the max numbers.
be2net incorrectly reports the queue used values for these fields.

Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: workaround to fix a bug in BE

BugLink: http://bugs.launchpad.net/bugs/921793

disable Tx vlan offloading in certain cases.

Signed-off-by: Ajit Khaparde <ajit.khaparde@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: update some counters to display via ethtool

BugLink: http://bugs.launchpad.net/bugs/921793

update pmem_fifo_overflow_drop, rx_priority_pause_frames counters.

Signed-off-by: Ajit Khaparde <ajit.khaparde@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: netpoll support

BugLink: http://bugs.launchpad.net/bugs/921793

Add missing netpoll support.

Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: Fix non utilization of RX queues

BugLink: http://bugs.launchpad.net/bugs/921793

When non power of two MSIX vectors are given to driver, some RX queues
are not utilized. Program RSS table in such a way that all queues
are utilized.

Signed-off-by: Padmanabh Ratnakar <padmanabh.ratnakar@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: Use new hash key

BugLink: http://bugs.launchpad.net/bugs/921793

This new hash key gives better distribution of packets across RX
queues.

Signed-off-by: Padmanabh Ratnakar <padmanabh.ratnakar@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: Add error handling for Lancer

BugLink: http://bugs.launchpad.net/bugs/921793

Detect error in Lancer by polling a HW register and
recover from this error if it is recoverable.

Signed-off-by: Padmanabh Ratnakar <padmanabh.ratnakar@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: Fix error recovery paths

BugLink: http://bugs.launchpad.net/bugs/921793

When TX queues are created again after error recovery,
netif_set_real_num_tx_queues() is invoked to update number of real
TX queues created. rtnl lock needs to be held when invoking this routine.

Signed-off-by: Padmanabh Ratnakar <padmanabh.ratnakar@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: Move to new SR-IOV implementation in Lancer

BugLink: http://bugs.launchpad.net/bugs/921793

SR-IOV implementation is Lancer has changed in following ways -
1)PF driver assigns one MAC addresses for VF using COMMON_SET_IFACE_MAC_LIST.
2)VF driver queries its MAC address using COMMON_GET_IFACE_MAC_LIST command
and assigns it to its interface.

Signed-off-by: Mammatha Edhala <mammatha.edhala@emulex.com>
Signed-off-by: Padmanabh Ratnakar <padmanabh.ratnakar@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: Use V1 query link status command for lancer

BugLink: http://bugs.launchpad.net/bugs/921793

Use V1 version of query link status command for Lancer.

Signed-off-by: Padmanabh Ratnakar <padmanabh.ratnakar@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: Fix VLAN promiscous mode for Lancer

BugLink: http://bugs.launchpad.net/bugs/921793

To enable VLAN promiscous mode, the HW interface should be created
with VLAN promiscous capability in Lancer. Add this capability during
creation of the HW interface.

Signed-off-by: Padmanabh Ratnakar <padmanabh.ratnakar@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: Add EEPROM dump feature for Lancer

BugLink: http://bugs.launchpad.net/bugs/921793

Implemented eeprom dump using ethtool feature for Lancer.

Signed-off-by: Padmanabh Ratnakar <padmanabh.ratnakar@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: add register dump feature for Lancer

BugLink: http://bugs.launchpad.net/bugs/921793

Implement register dump using ethtool for Lancer.

Signed-off-by: Padmanabh Ratnakar <padmanabh.ratnakar@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: Fix TX queue create for Lancer

BugLink: http://bugs.launchpad.net/bugs/921793

Lancer uses V1 version of TXQ create. This command needs interface
id for TX queue creation. Rearrange code such that tx queue create
is after interface create. As TXQ create is now called after MCC
ring create use MCC instead of MBOX.

Signed-off-by: Padmanabh Ratnakar <padmanabh.ratnakar@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: stop issuing FW cmds if any cmd times out

BugLink: http://bugs.launchpad.net/bugs/921793

A FW cmd timeout (with a sufficiently large timeout value in the
order of tens of seconds) indicates an unresponsive FW. In this state
issuing further cmds and waiting for a completion will only stall the process.

Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: don't log more than one error on detecting EEH/UE errors

BugLink: http://bugs.launchpad.net/bugs/921793

Currently we're spamming error messages each time a FW cmd call is made
while in EEH/UE error state. One log msg on error detection is enough.

Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: stop checking the UE registers after an EEH error

BugLink: http://bugs.launchpad.net/bugs/921793

Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

be2net: init (vf)_if_handle/vf_pmac_id to handle failure scenarios

BugLink: http://bugs.launchpad.net/bugs/921793

Initialize if_handle, vf_if_handle and vf_pmac_id with "-1" so that in
failure cases when be_clear() is called, we can skip over
if_destroy/pmac_del cmds if they have not been created.

Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

sweep the floors and convert some .get_drvinfo routines to strlcpy

BugLink: http://bugs.launchpad.net/bugs/921793

Per the mention made by Ben Hutchings that strlcpy is now the preferred
string copy routine for a .get_drvinfo routine, do a bit of floor
sweeping and convert some of the as-yet unconverted ethernet drivers to
it.

Signed-off-by: Rick Jones <rick.jones2@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

UBUNTU: SAUCE: remove __initdata from vesafb_fix

BugLink: http://bugs.launchpad.net/bugs/969309

OK. Then, I think we also want to fix these warnings probably introduced by
commit a6021559 "UBUNTU: SAUCE: (no-up) Modularize vesafb".

WARNING: drivers/video/vesafb.o(.exit.text+0x42): Section mismatch in reference from the function vesafb_remove() to the (unknown reference) .init.data:(unknown)
The function __exit vesafb_remove() references
a (unknown reference) __initdata (unknown).
This is often seen when error handling in the exit function
uses functionality in the init path.
The fix is often to remove the __initdata annotation of
(unknown) so it may be used outside an init section.

WARNING: drivers/video/vesafb.o(.exit.text+0x4a): Section mismatch in reference from the function vesafb_remove() to the variable .init.data:vesafb_fix
The function __exit vesafb_remove() references
a variable __initdata vesafb_fix.
This is often seen when error handling in the exit function
uses functionality in the init path.
The fix is often to remove the __initdata annotation of
vesafb_fix so it may be used outside an init section.

Reported-by: Tetsuo Honda <from-ubuntu@I-love.SAKURA.ne.jp>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

Revert "UBUNTU: SAUCE: hotkey quirks for various Zeptro Znote and Fujitsu Amilo laptops"

This reverts commit f5931d1fc7c699a650ca14a31f4b4bc620977389.

These have already been converted to udev rules so we no longer need
these to be quirked in the kernel.

Signed-off-by: Andy Whitcroft <apw@canonical.com>

UBUNTU: d-i: Fix module name for dm-raid45

This was incorrectly named in the md-modules section.

BugLink: http://bugs.launchpad.net/bugs/969248

Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

UBUNTU: SAUCE: PCI: Allow pcie_aspm=force to work even when FADT indicates it is unsupported

Submitted upstream.
BugLink: http://bugs.launchpad.net/bugs/962038

Right now using pcie_aspm=force will not enable ASPM if the FADT indicates
ASPM is unsupported.  However, the semantics of force should probably allow
for this, especially as they did before the ASPM disable rework with commit
3c076351c4027a56d5005a39a0b518a4ba393ce2

This patch just skips the clearing of any ASPM setup that the firmware has
carried out on this bus if pcie_aspm=force is being used.

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Acked-by: Stefan Bader <stefan.bader@canonical.com>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

UBUNTU: SAUCE (drop after 3.4) Quirk for enabling backlight hotkeys on Samsung N150P

BugLink: http://bugs.launchpad.net/bugs/875893

This patch is a very small part of a much larger patch that is upstream. I felt
the full patch was too invasive to try to backport.

(partial backport of upstream commit: a979e2e2af7d5b4bb3b20f6a716c627bb23a6753)
Signed-off-by: Brad Figg <brad.figg@canonical.com>
Acked-by: Seth Forshee <seth.forshee@canonical.com>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

ata_piix: Add Toshiba Satellite Pro A120 to the quirks list due to broken suspend functionality.

BugLink: http://bugs.launchpad.net/bugs/886850

Signed-off-by: Benjamin Larsson <benjamin@southpole.se>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
(cherry picked from commit b73fa4630e63e4d23407181c2bc0a54777bd0ce8)
Signed-off-by: Brad Figg <brad.figg@canonical.com>
Signed-off-by: Leann Ogasawara <leann.ogasawara@canonical.com>

toshiba_acpi: Refuse to load on machines with buggy INFO implementations

BugLink: http://bugs.launchpad.net/bugs/810015

Several Satellite models have a buggy implementation of the INFO method
that causes ACPI exceptions when executed:

 ACPI Error: Result stack is empty! State=ffff88012d70f800 (20110413/dswstate-98)
 ACPI Exception: AE_AML_NO_RETURN_VALUE, Missing or null operand (20110413/dsutils-646)
 ACPI Exception: AE_AML_NO_RETURN_VALUE, While creating Arg 0 (20110413/dsutils-763)
 ACPI Error: Method parse/execution failed [\_SB_.VALZ.GETE] (Node ffff880131175eb0), AE_AML_NO_RETURN_VALUE (20110413/psparse-536)
 ACPI Error: Method parse/execution failed [\_SB_.VALZ.INFO] (Node ffff880131175ed8), AE_AML_NO_RETURN_VALUE (20110413/psparse-536)
 toshiba_acpi: ACPI INFO method execution failed
 toshiba_acpi: Failed to query hotkey event

All known machines with this implementation also have a WMI interface
with event GUID 59142400-C6A3-40FA-BADB-8A2652834100 which is not seen
on any other models. Refuse to load toshiba_acpi on machines with this
guid.

Signed-off-by: Seth Forshee <seth.forshee@canonical.com>
Signed-off-by: Matthew Garrett <mjg@redhat.com>
(cherry picked from commit f11f999e989061952f1a27bd0c49645a46d13173)

Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

toshiba_acpi: Support additional hotkey scancodes

BugLink: http://bugs.launchpad.net/bugs/810015

These scancodes are used by many of the models now supported with
the addition of TOS1900 device support.

Signed-off-by: Azael Avalos <coproscefalo@gmail.com>
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>
Signed-off-by: Matthew Garrett <mjg@redhat.com>
(cherry picked from commit af502837a08c8ca3495d3940d4c4eb4e19a3beaa)

Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

toshiba_acpi: Support alternate hotkey interfaces

BugLink: http://bugs.launchpad.net/bugs/810015

There are two types of problems that prevent hotkeys from working
on many of the machines supported by toshiba_acpi. The first of
these is the lack of a functioning SCI for hotkey events. For these
machines it is possible to filter the Fn keypresses from the
keyboard and generate a notification by executing the ACPI NTFY
method.

The second problem is a lack of support for HCI_SYSTEM_EVENT, which
is used for reading the hotkey scancodes. On these machines the
scancodes can be read by executing the ACPI NTFY method.

This patch fixes both problems by installing an i8042 filter when
the NTFY method is present to generate notifications and by
detecting which of INFO or HCI_SYSTEM_EVENT is supported for
reading scancodes. If neither method of reading scancodes is
supported, the hotkey input device is not registered.

Signed-off-by: Azael Avalos <coproscefalo@gmail.com>
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>
Signed-off-by: Matthew Garrett <mjg@redhat.com>
(cherry picked from commit 29cd293f9f8cd76444657622980010b9364b1de6)

Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

ACPI: EC: Add ec_get_handle()

BugLink: http://bugs.launchpad.net/bugs/810015

toshiba_acpi needs to execute an AML method within the EC namespace
to make hotkeys work on some platforms. Provide an interface to
allow it to easily get a handle to the EC namespace for this purpose.

Signed-off-by: Seth Forshee <seth.forshee@canonical.com>
Signed-off-by: Matthew Garrett <mjg@redhat.com>
(cherry picked from commit 3e2abc5a35d25442821e1733687b7abbc83b5072)

Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

toshiba_acpi: make one-bit bitfields unsigned

BugLink: http://bugs.launchpad.net/bugs/810015

This doesn't change how the code works, but it silences a Sparse
complaint:
drivers/platform/x86/toshiba_acpi.c:121:37: error: dubious one-bit signed bitfield

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Matthew Garrett <mjg@redhat.com>
(cherry picked from commit 592b746c55b63770da25148ea3b56ed463a596b2)

Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>

UBUNTU: SAUCE (no-up) Provide a param for allowing the BIOS to handle changing the brightness on AC/battery status changes.

BugLink: http://bugs.launchpad.net/bugs/949311

We currently carry a SAUCE patch which lets the OS handle the brightness
levels automatically when connecting/disconnecting AC. There are some
laptops (MSI Wind) for which this doesn't work. Provide a driver param
which allows this behaviour to be overriden.

Signed-off-by: Brad Figg <brad.figg@canonical.com>
Acked-by: Colin King <colin.king@canonical.com>
Signed-off-by: Leann Ogasawara <leann.ogasawara@canonical.com>

UBUNTU: Bump ABI

Ignore: yes
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

UBUNTU: Start new release

Ignore: yes
Signed-off-by: Leann Ogasawara <leann.ogasawara@canonical.com>

UBUNTU: Ubuntu-3.2.0-21.34

Signed-off-by: Leann Ogasawara <leann.ogasawara@canonical.com>

UBUNTU: SAUCE: disable_nx should not be in __cpuinitdata section for X86_32

BugLink: http://bugs.launchpad.net/bugs/968233

I noticed a section mismatch warning while building 3.2.0-20.33 for X86_32.

 AR      arch/x86/lib/lib.a
  LD      vmlinux.o
  MODPOST vmlinux.o
WARNING: vmlinux.o(.text+0x187833): Section mismatch in reference from the function load_elf_binary() to the variable .cpuinit.data:disable_nx
The function load_elf_binary() references
the variable __cpuinitdata disable_nx.
This is often because load_elf_binary lacks a __cpuinitdata
annotation or the annotation of disable_nx is wrong.

load_elf_binary() is definitely called after initialization.

This code was added by 'UBUNTU: ubuntu: nx-emu - i386: NX emulation', so
this is not an upstream problem.

Reported-by: Tetsuo Handa <from-ubuntu@I-love.SAKURA.ne.jp>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

UBUNTU: d-i: Add brcm*mac drivers to nic-modules

Signed-off-by: Leann Ogasawara <leann.ogasawara@canonical.com>

UBUNTU: d-i: Remove obsolete bcm43xx* drivers from nic-modules

Signed-off-by: Leann Ogasawara <leann.ogasawara@canonical.com>

UBUNTU: d-i: Add missing iwl* drivers to nic-modules

BugLink: http://bugs.launchpad.net/bugs/965116

Signed-off-by: Leann Ogasawara <leann.ogasawara@canonical.com>

UBUNTU: d-i: Fix iwl3945 driver name typo in nic-modules

s/ipw3945/iwl3945/ in nic-modules file

Signed-off-by: Leann Ogasawara <leann.ogasawara@canonical.com>

drm/i915: Add Clientron E830 to the ignore LVDS list

BugLink: http://bugs.launchpad.net/bugs/907055

Signed-off-by: Joel Sass <jsass@disklessworkstations.com>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
(cherry picked from commit 44306ab302687b519a31aa498b954c1e26f95a6b)

Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

drm/i915: Ignore LVDS on hp t5745 and hp st5747 thin client

Add a no_lvds quirk for the HP t5745 and HP st5747 thin clients

dmidecode for those thin clients are attached in thoses bugs:
BugLink: https://bugs.launchpad.net/bugs/911916
BugLink: https://bugs.launchpad.net/bugs/911920

Signed-off-by: Marc Gariepy <mgariepy@ubuntu.com>
Acked-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
(cherry picked from commit f5b8a7ed0405d48fd096acce48fbefbed77fb055)

Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

UBUNTU: SAUCE: AppArmor: Fix change_onexec when called from a confined task

Patch submitted upstream, but not merged yet.

Fix failure in aa_change_onexec api when the request is made from a confined
task.  This failure was caused by two problems

 The AA_MAY_ONEXEC perm was not being mapped correctly for this case.

 The executable name was being checked as second time instead of using the
 requested onexec profile name, which may not be the same as the exec
 profile name. This mistake can not be exploited to grant extra permission
 because of the above flaw where the ONEXEC permission was not being mapped
 so it will not be granted.

BugLink: http://bugs.launchpad.net/bugs/963756

Signed-off-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

x86, tsc: Fix SMI induced variation in quick_pit_calibrate()

BugLink: http://bugs.launchpad.net/bugs/965586

pit_expect_msb() returns success wrongly in the below SMI scenario:

a. pit_verify_msb() has not yet seen the MSB transition.

b. we are close to the MSB transition though and got a SMI immediately after
   returning from pit_verify_msb() which didn't see the MSB transition. PIT MSB
   transition has happened somewhere during SMI execution.

c. returned from SMI and we noted down the 'tsc', saw the pit MSB change now and
   exited the loop to calculate 'deltatsc'. Instead of noting the TSC at the MSB
   transition, we are way off because of the SMI.  And as the SMI happened
   between the pit_verify_msb() and before the 'tsc' is recorded in the
   for loop, 'delattsc' (d1/d2 in quick_pit_calibrate()) will be small and
   quick_pit_calibrate() will not notice this error.

Depending on whether SMI disturbance happens while computing d1 or d2, we will
see the TSC calibrated value smaller or bigger than the expected value. As a
result, in a cluster we were seeing a variation of approximately +/- 20MHz in
the calibrated values, resulting in NTP failures.

  [ As far as the SMI source is concerned, this is a periodic SMI that gets
    disabled after ACPI is enabled by the OS. But the TSC calibration happens
    before the ACPI is enabled. ]

To address this, change pit_expect_msb() so that

 - the 'tsc' is the TSC in between the two reads that read the MSB
change from the PIT (same as before)

 - the 'delta' is the difference in TSC from *before* the MSB changed
to *after* the MSB changed.

Now the delta is twice as big as before (it covers four PIT accesses,
roughly 4us) and quick_pit_calibrate() will loop a bit longer to get
the calibrated value with in the 500ppm precision. As the delta (d1/d2)
covers four PIT accesses, actual calibrated result might be closer to
250ppm precision.

As the loop now takes longer to stabilize, double MAX_QUICK_PIT_MS to 50.

SMI disturbance will showup as much larger delta's and the loop will take
longer than usual for the result to be with in the accepted precision. Or will
fallback to slow PIT calibration if it takes more than 50msec.

Also while we are at this, remove the calibration correction that aims to
get the result to the middle of the error bars. We really don't know which
direction to correct into, so remove it.

Reported-and-tested-by: Suresh Siddha <suresh.b.siddha@intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Link: http://lkml.kernel.org/r/1326843337.5291.4.camel@sbsiddha-mobl2
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
(cherry picked from commit 68f30fbee19cc67849b9fa8e153ede70758afe81)

Signed-off-by: Leann Ogasawara <leann.ogasawara@canonical.com>

UBUNTU: SAUCE: fsam7400: use UMH_WAIT_PROC consistently

BugLink: http://bugs.launchpad.net/bugs/963685

Inspired by Testuo Handa, "usermodehelper: use UMH_WAIT_PROC consistently"

Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

kmod: make __request_module() killable

BugLink: http://bugs.launchpad.net/bugs/963685

As Tetsuo Handa pointed out, request_module() can stress the system
while the oom-killed caller sleeps in TASK_UNINTERRUPTIBLE.

The task T uses "almost all" memory, then it does something which
triggers request_module().  Say, it can simply call sys_socket().  This
in turn needs more memory and leads to OOM.  oom-killer correctly
chooses T and kills it, but this can't help because it sleeps in
TASK_UNINTERRUPTIBLE and after that oom-killer becomes "disabled" by the
TIF_MEMDIE task T.

Make __request_module() killable.  The only necessary change is that
call_modprobe() should kmalloc argv and module_name, they can't live in
the stack if we use UMH_KILLABLE.  This memory is freed via
call_usermodehelper_freeinfo()->cleanup.

Reported-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Tejun Heo <tj@kernel.org>
Cc: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
(cherry picked from commit 1cc684ab75123efe7ff446eb821d44375ba8fa30)

Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

kmod: introduce call_modprobe() helper

BugLink: http://bugs.launchpad.net/bugs/963685

No functional changes.  Move the call_usermodehelper code from
__request_module() into the new simple helper, call_modprobe().

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Tejun Heo <tj@kernel.org>
Cc: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
(cherry picked from commit 3e63a93b987685f02421e18b2aa452d20553a88b)

Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

usermodehelper: ____call_usermodehelper() doesn't need do_exit()

BugLink: http://bugs.launchpad.net/bugs/963685

Minor cleanup.  ____call_usermodehelper() can simply return, no need to
call do_exit() explicitely.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Tejun Heo <tj@kernel.org>
Cc: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
(cherry picked from commit 5b9bd473e3b8a8c6c4ae99be475e6e9b27568555)

Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

usermodehelper: kill umh_wait, renumber UMH_* constants

BugLink: http://bugs.launchpad.net/bugs/963685

No functional changes.  It is not sane to use UMH_KILLABLE with enum
umh_wait, but obviously we do not want another argument in
call_usermodehelper_* helpers.  Kill this enum, use the plain int.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Tejun Heo <tj@kernel.org>
Cc: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
(cherry picked from commit 9d944ef32e83405a07376f112e9f02161d3e9731)

Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

usermodehelper: implement UMH_KILLABLE

BugLink: http://bugs.launchpad.net/bugs/963685

Implement UMH_KILLABLE, should be used along with UMH_WAIT_EXEC/PROC.
The caller must ensure that subprocess_info->path/etc can not go away
until call_usermodehelper_freeinfo().

call_usermodehelper_exec(UMH_KILLABLE) does
wait_for_completion_killable.  If it fails, it uses
xchg(&sub_info->complete, NULL) to serialize with umh_complete() which
does the same xhcg() to access sub_info->complete.

If call_usermodehelper_exec wins, it can safely return.  umh_complete()
should get NULL and call call_usermodehelper_freeinfo().

Otherwise we know that umh_complete() was already called, in this case
call_usermodehelper_exec() falls back to wait_for_completion() which
should succeed "very soon".

Note: UMH_NO_WAIT == -1 but it obviously should not be used with
UMH_KILLABLE.  We delay the neccessary cleanup to simplify the back
porting.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Tejun Heo <tj@kernel.org>
Cc: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
(cherry picked from commit d0bd587a80960d7ba7e0c8396e154028c9045c54)

Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

usermodehelper: introduce umh_complete(sub_info)

BugLink: http://bugs.launchpad.net/bugs/963685

Preparation.  Add the new trivial helper, umh_complete().  Currently it
simply does complete(sub_info->complete).

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Tejun Heo <tj@kernel.org>
Cc: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
(cherry picked from commit b3449922502f5a161ee2b5022a33aec8472fbf18)

Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

usermodehelper: use UMH_WAIT_PROC consistently

BugLink: http://bugs.launchpad.net/bugs/963685

A few call_usermodehelper() callers use the hardcoded constant instead of
the proper UMH_WAIT_PROC, fix them.

Reported-by: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Lars Ellenberg <drbd-dev@lists.linbit.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Michal Januszewski <spock@gentoo.org>
Cc: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Cc: Kentaro Takeda <takedakn@nttdata.co.jp>
Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: James Morris <jmorris@namei.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
(cherry picked from commit 70834d3070c3f3015ab5c05176d54bd4a0100546)

Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

UBUNTU: Rebase to v3.2.13

Signed-off-by: Leann Ogasawara <leann.ogasawara@canonical.com>

UBUNTU: [Config]: enable CONFIG_DRM_LOAD_EDID_FIRMWARE

Signed-off-by: Andy Whitcroft <apw@canonical.com>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

drm: allow loading an EDID as firmware to override broken monitor

Broken monitors and/or broken graphic boards may send erroneous or no
EDID data. This also applies to broken KVM devices that are unable to
correctly forward the EDID data of the connected monitor but invent
their own fantasy data.

This patch allows to specify an EDID data set to be used instead of
probing the monitor for it. It contains built-in data sets of frequently
used screen resolutions. In addition, a particular EDID data set may be
provided in the /lib/firmware directory and loaded via the firmware
interface. The name is passed to the kernel as module parameter of the
drm_kms_helper module either when loaded
  options drm_kms_helper edid_firmware=edid/1280x1024.bin
or as kernel commandline parameter
  drm_kms_helper.edid_firmware=edid/1280x1024.bin

It is also possible to restrict the usage of a specified EDID data set
to a particular connector. This is done by prepending the name of the
connector to the name of the EDID data set using the syntax
  edid_firmware=[<connector>:]<edid>
such as, for example,
  edid_firmware=DVI-I-1:edid/1920x1080.bin
in which case no other connector will be affected.

The built-in data sets are
Resolution    Name
--------------------------------
1024x768      edid/1024x768.bin
1280x1024     edid/1280x1024.bin
1680x1050     edid/1680x1050.bin
1920x1080     edid/1920x1080.bin

They are ignored, if a file with the same name is available in the
/lib/firmware directory.

The built-in EDID data sets are based on standard timings that may not
apply to a particular monitor and even crash it. Ideally, EDID data of
the connected monitor should be used. They may be obtained through the
drm/cardX/cardX-<connector>/edid entry in the /sys/devices PCI directory
of a correctly working graphics adapter.

It is even possible to specify the name of an EDID data set on-the-fly
via the /sys/module interface, e.g.
echo edid/myedid.bin >/sys/module/drm_kms_helper/parameters/edid_firmware
The new screen mode is considered when the related kernel function is
called for the first time after the change. Such calls are made when the
X server is started or when the display settings dialog is opened in an
already running X server.

Signed-off-by: Carsten Emde <C.Emde@osadl.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>

(cherry-picked from da0df92b57311aa1b26a2a90599ed16e1e968b90 drm-next)
Signed-off-by: Andy Whitcroft <apw@canonical.com>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

UBUNTU: Add isci to debian scsi-modules udeb

BugLink: http://bugs.launchpad.net/bugs/963306

Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

UBUNTU: SAUCE: Update aufs for build failure caused by apparmor backport

Fix build failure in aufs introduced by

  commit 9cd98c046b57cd1bdbd53c3669f6cdd75edffd61

which has been backported from 3.4 as part of the AppArmor 3.4 backport

Signed-off-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

UBUNTU: SAUCE: AppArmor: basic networking rules

Base support for network mediation.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

UBUNTU: SAUCE: AppArmor: Add profile introspection file to interface

Add the dynamic profiles file to the interace, to allow load policy
introspection.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Kees Cook <kees@ubuntu.com>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

UBUNTU: SAUCE: AppArmor: Add the ability to mediate mount

Add the ability for apparmor to do mediation of mount operations. Mount
rules require an updated apparmor_parser (2.8 series) for policy compilation.

The basic form of the rules are.

  [audit] [deny] mount [conds]* [device] [ -> [conds] path],
  [audit] [deny] remount [conds]* [path],
  [audit] [deny] umount [conds]* [path],
  [audit] [deny] pivotroot [oldroot=<value>] <path>

  remount is just a short cut for mount options=remount

  where [conds] can be
    fstype=<expr>
    options=<expr>

Example mount commands
  mount, # allow all mounts, but not umount or pivotroot

  mount fstype=procfs,  # allow mounting procfs anywhere

  mount options=(bind, ro) /foo -> /bar,  # readonly bind mount

  mount /dev/sda -> /mnt,

  mount /dev/sd** -> /mnt/**,

  mount fstype=overlayfs options=(rw,upperdir=/tmp/upper/,lowerdir=/) -> /mnt/

  umount,

  umount /m*,

See the apparmor userspace for full documentation

Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Kees Cook <kees@ubuntu.com>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

AppArmor: Fix location of const qualifier on generated string tables

Signed-off-by: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit 7e570145cb022beeb58e3f691e0418477b670223)
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

AppArmor: add const qualifiers to string arrays

Signed-off-by: Jan Engelhardt <jengelh@medozas.de>
Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit 2d4cee7e3a2b9f9c3237672cc136e20dbad0e2ce)
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

AppArmor: Add ability to load extended policy

Add the base support for the new policy extensions. This does not bring
any additional functionality, or change current semantics.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Kees Cook <kees@ubuntu.com>
(cherry picked from commit ad5ff3db53c68c2f12936bc74ea5dfe0af943592)
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

AppArmor: Move path failure information into aa_get_name and rename

Move the path name lookup failure messages into the main path name lookup
routine, as the information is useful in more than just aa_path_perm.

Also rename aa_get_name to aa_path_name as it is not getting a reference
counted object with a corresponding put fn.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Kees Cook <kees@ubuntu.com>
(cherry picked from commit 57fa1e18091e66b7e1002816523cb218196a882e)
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

AppArmor: Update dfa matching routines.

Update aa_dfa_match so that it doesn't result in an input string being
walked twice (once to get its length and another time to match)

Add a single step functions
  aa_dfa_next

Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Kees Cook <kees@ubuntu.com>
(cherry picked from commit 0fe1212d0539eb6c1e27d388711172d786e299cc)
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

AppArmor: Minor cleanup of d_namespace_path to consolidate error handling

Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Kees Cook <kees@ubuntu.com>
(cherry picked from commit 3372b68a3c982611dcc30b3c872f8bbdee019e5e)
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

AppArmor: Retrieve the dentry_path for error reporting when path lookup fails

When __d_path and d_absolute_path fail due to the name being outside of
the current namespace no name is reported.  Use dentry_path to provide
some hint as to which file was being accessed.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Kees Cook <kees@ubuntu.com>
(cherry picked from commit fbba8d89acea5d628d1d076b1d8962db438ff832)
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

AppArmor: Add const qualifiers to generated string tables

Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit 33e521acff709d275950ec5bf8dd577d873cd61e)
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

AppArmor: Fix oops in policy unpack auditing

Post unpacking of policy a verification pass is made on x transition
indexes.  When this fails a call to audit_iface is made resulting in an
oops, because audit_iface is expecting a valid buffer position but
since the failure comes from post unpack verification there is none.

Make the position argument optional so that audit_iface can be called
from post unpack verification.

Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit b1b4bc2ed94d157f3ed60c17a12b658ccb96a76f)
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

AppArmor: Fix error returned when a path lookup is disconnected

The returning of -ESATLE when a path lookup fails as disconnected is wrong.
Since AppArmor is rejecting the access return -EACCES instead.

This also fixes a bug in complain (learning) mode where disconnected paths
are denied because -ESTALE errors are not ignored causing failures that
can change application behavior.

Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit ef9a762279c9ce98c592fb144b31898411feb94d)
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

AppArmor: Fix the error case for chroot relative path name lookup

When a chroot relative pathname lookup fails it is falling through to
do a d_absolute_path lookup.  This is incorrect as d_absolute_path should
only be used to lookup names for namespace absolute paths.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Kees Cook <kees@ubuntu.com>
(cherry picked from commit 28042fabf43b9a8ccfaa38f8c8187cc525e53fd3)
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

AppArmor: fix mapping of META_READ to audit and quiet flags

The mapping of AA_MAY_META_READ for the allow mask was also being mapped
to the audit and quiet masks. This would result in some operations being
audited when the should not.

This flaw was hidden by the previous audit bug which would drop some
messages that where supposed to be audited.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Kees Cook <kees@ubuntu.com>
(cherry picked from commit 38305a4bab4be5d278443b057f7f5e97afb07f26)
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

AppArmor: Fix underflow in xindex calculation

If the xindex value stored in the accept tables is 0, the extraction of
that value will result in an underflow (0 - 4).

In properly compiled policy this should not happen for file rules but
it may be possible for other rule types in the future.

To exploit this underflow a user would have to be able to load a corrupt
policy, which requires CAP_MAC_ADMIN, overwrite system policy in kernel
memory or know of a compiler error resulting in the flaw being present
for loaded policy (no such flaw is known at this time).

Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Kees Cook <kees@ubuntu.com>
(cherry picked from commit 8b964eae204d791421677ec56b94a7b18cf8740d)
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

AppArmor: Fix dropping of allowed operations that are force audited

The audit permission flag, that specifies an audit message should be
provided when an operation is allowed, was being ignored in some cases.

This is because the auto audit mode (which determines the audit mode from
system flags) was incorrectly assigned the same value as audit mode. The
shared value would result in messages that should be audited going through
a second evaluation as to whether they should be audited based on the
auto audit, resulting in some messages being dropped.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Kees Cook <kees@ubuntu.com>
(cherry picked from commit ade3ddc01e2e426cc24c744be85dcaad4e8f8aba)
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>

AppArmor: Add mising end of structure test to caps unpacking

The unpacking of struct capsx is missing a check for the end of the
caps structure.  This can lead to unpack failures depending on what else
is packed into the policy file being unpacked.

Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Kees Cook <kees@ubuntu.com>
(cherry picked from commit cdbd2884df8ad026143bb482a96d38e616947b17)
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>