pandora-kernel.git
13 years agowl1271: Fix TX starvation
Ido Yariv [Tue, 12 Oct 2010 12:49:10 +0000 (14:49 +0200)]
wl1271: Fix TX starvation

While wl1271_irq_work handles RX directly (by calling wl1271_rx), a different
work is scheduled for transmitting packets. The IRQ work might handle more than
one interrupt during a single call, including multiple TX completion
interrupts. This might starve TX, since no packets are transmitted until all
interrupts are handled.

Fix this by calling the TX work function directly, instead of deferring
it.

Signed-off-by: Ido Yariv <ido@wizery.com>
Reviewed-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
13 years agowl1271: TX aggregation optimization
Ido Yariv [Tue, 12 Oct 2010 12:49:09 +0000 (14:49 +0200)]
wl1271: TX aggregation optimization

In case the aggregation buffer is too small to hold all available packets,
the buffer is transferred to the FW and no more packets are aggregated.
Although there may be enough available TX blocks, no additional packets will
be handled by the current TX work.

Fix this by flushing the aggregation buffer when it's full, and continue
transferring packets as long as there are enough available TX blocks.

Signed-off-by: Ido Yariv <ido@wizery.com>
Reviewed-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
13 years agowl1271: ref_clock cosmetic changes
Gery Kahn [Tue, 5 Oct 2010 14:09:05 +0000 (16:09 +0200)]
wl1271: ref_clock cosmetic changes

Cosmetic cleanup for ref_clock code while configured by board.

Signed-off-by: Gery Kahn <geryk@ti.com>
Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
13 years agoiwlagn: use 6000g2b uCode for 130 series devices
Wey-Yi Guy [Sat, 23 Oct 2010 16:15:46 +0000 (09:15 -0700)]
iwlagn: use 6000g2b uCode for 130 series devices

For 130 series device, 6000g2b uCode will be used, no need to
have additional defines for 130 devices, so remove those.

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoiwlwifi: add new devices to Kconfig
Wey-Yi Guy [Sat, 23 Oct 2010 16:15:45 +0000 (09:15 -0700)]
iwlwifi: add new devices to Kconfig

Adding description to Kconfig to indicate more devices
are being supported by iwlagn

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoiwlagn: turn dynamic smps on while BT is on
Wey-Yi Guy [Sat, 23 Oct 2010 16:15:44 +0000 (09:15 -0700)]
iwlagn: turn dynamic smps on while BT is on

While BT is on and doing iscan and/or pscan, BT is in listen
mode which will impact WiFi throughput, we need to
enable dynamic smps in order to improve the rx throughput.

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoiwlwifi: implement switching iftype while up
Johannes Berg [Sat, 23 Oct 2010 16:15:43 +0000 (09:15 -0700)]
iwlwifi: implement switching iftype while up

Implement switching the interface while an
interface is up in iwlwifi. Interfaces have
to stay on the context they were created on.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoiwlwifi: fix IBSS beaconing
Johannes Berg [Sat, 23 Oct 2010 16:15:42 +0000 (09:15 -0700)]
iwlwifi: fix IBSS beaconing

My previous patch to clean up all RXON handling
inadvertently broke IBSS because it failed to
take into account that unlike in AP mode, IBSS
requires beacons to be sent only after setting
the RXON assoc.

Fix this, clean up the code a bit, improve the
error checking around this, and also react to
beacon changes in IBSS mode from mac80211.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoiwlagn: new RXON processing for modern devices
Johannes Berg [Sat, 23 Oct 2010 16:15:41 +0000 (09:15 -0700)]
iwlagn: new RXON processing for modern devices

In order to simplify the flow, and make new
enhancements easier, separate out the RXON
processing for modern AGN (5000 and newer)
from RXON processing for the older 3945 and
4965 devices. Avoid changing these old ones
to avoid regressions and move their code to
a new file (iwl-legacy.c). 4965 gets the
commit_rxon that used to be common for all
AGN devices, but with removed PAN support.

The new RXON processing is more central and
does more work in committing, so that it is
easier to follow.

To make it more evident what is split out
for legacy, split the necessary operations
for that into a new struct iwl_legacy_ops.
Those parts that still exist in the new AGN
code don't need to be parametrized.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoiwlagn: don't resend RXON timing
Johannes Berg [Sat, 23 Oct 2010 16:15:40 +0000 (09:15 -0700)]
iwlagn: don't resend RXON timing

Resending RXON timing here caused issues with
dual-mode under certain circumstances, so avoid
doing it here right now.

This effectively reverts b01efe434bd6ea807eb72b
and partially 2491fa42d9bdf26075765. The next
patch will make all this cleaner for just the
devices that need it.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoiwlwifi: make mac80211 ops a device config
Johannes Berg [Sat, 23 Oct 2010 16:15:39 +0000 (09:15 -0700)]
iwlwifi: make mac80211 ops a device config

In the future, 4965 and modern AGN devices will
need to have different mac80211 callbacks since
they have different capabilities. Prepare for
that by making the mac80211 operations a device
config.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: refactor debugfs function generation code
Eliad Peller [Wed, 27 Oct 2010 12:58:29 +0000 (14:58 +0200)]
mac80211: refactor debugfs function generation code

refactor mac80211 debugfs code by using a format&copy function, instead of
duplicating the code for each generated function.

this change reduces about 600B from mac80211.ko

Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: content DMA start / stop through the PCU lock
Luis R. Rodriguez [Tue, 26 Oct 2010 22:27:25 +0000 (15:27 -0700)]
ath9k: content DMA start / stop through the PCU lock

This helps align resets / RX enable & disable / TX stop / start.
Locking around the PCU is important to ensure the hardware doesn't
get stale data when working with DMA'able data.

This is part of a series of patches which fix stopping
TX DMA completley when requested on the driver.
For more details about this issue refer to this thread:

http://marc.info/?l=linux-wireless&m=128629803703756&w=2

Tested-by: Ben Greear <greearb@candelatech.com>
Cc: Kyungwan Nam <kyungwan.nam@atheros.com>
Cc: stable@kernel.org
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: move the PCU lock to the sc structure
Luis R. Rodriguez [Tue, 26 Oct 2010 22:27:24 +0000 (15:27 -0700)]
ath9k: move the PCU lock to the sc structure

The PCU lock should be used to contend TX DMA as well,
this will be done next.

This is part of a series of patches which fix stopping
TX DMA completley when requested on the driver.
For more details about this issue refer to this thread:

http://marc.info/?l=linux-wireless&m=128629803703756&w=2

Tested-by: Ben Greear <greearb@candelatech.com>
Cc: Kyungwan Nam <kyungwan.nam@atheros.com>
Cc: stable@kernel.org
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: simplify hw reset locking
Luis R. Rodriguez [Tue, 26 Oct 2010 22:27:23 +0000 (15:27 -0700)]
ath9k: simplify hw reset locking

The new PCU lock is better placed so we can just contend
against that when trying to reset hardware.

This is part of a series of patches which fix stopping
TX DMA completley when requested on the driver.
For more details about this issue refer to this thread:

http://marc.info/?l=linux-wireless&m=128629803703756&w=2

Tested-by: Ben Greear <greearb@candelatech.com>
Cc: Kyungwan Nam <kyungwan.nam@atheros.com>
Cc: stable@kernel.org
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: Properly assign boolean types
Mohammed Shafi Shajakhan [Tue, 26 Oct 2010 15:58:57 +0000 (21:28 +0530)]
ath9k: Properly assign boolean types

This takes care that boolean types are properly assigned

Signed-off-by: Mohammed Shafi Shajakhan <mshajakhan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_htc: Handle monitor mode properly for HTC devices
Rajkumar Manoharan [Tue, 26 Oct 2010 14:41:29 +0000 (20:11 +0530)]
ath9k_htc: Handle monitor mode properly for HTC devices

No need to inform about monitor interface changes to
firmware. Set the HW mode to monitor type based on mac80211
indication flag is sufficient.

Signed-off-by: Rajkumar Manoharan <rmanoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoiwlwifi: Convert to new PCI PM framework
John W. Linville [Mon, 25 Oct 2010 20:12:37 +0000 (16:12 -0400)]
iwlwifi: Convert to new PCI PM framework

Use the new PCI PM and let the PCI core code handle the PCI-specific
details of power transitions.

Based on similarly titled ath9k patch posted by Rafael J. Wysocki.

Signed-off-by: John W. Linville <linville@tuxdriver.com>
Acked-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agort2x00pci: do not use GFP_DMA
John W. Linville [Mon, 25 Oct 2010 14:33:07 +0000 (10:33 -0400)]
rt2x00pci: do not use GFP_DMA

Signed-off-by: John W. Linville <linville@tuxdriver.com>
Acked-by: Ivo van Doorn <IvDoorn@gmail.com>
13 years agob43: remove extraneous code in free_ringmemory
John W. Linville [Mon, 25 Oct 2010 14:24:09 +0000 (10:24 -0400)]
b43: remove extraneous code in free_ringmemory

This code seems to have been cut-n-pasted from alloc_ringmemory?
Anyway, it is useless.

Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agocarl9170: fix typos
Hauke Mehrtens [Sat, 23 Oct 2010 17:51:32 +0000 (19:51 +0200)]
carl9170: fix typos

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Acked-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agob43: N-PHY: improve 2055 radio initialization
Rafał Miłecki [Fri, 22 Oct 2010 15:43:49 +0000 (17:43 +0200)]
b43: N-PHY: improve 2055 radio initialization

1) Upload 5 GHz values when needed.
2) Do not upload all values on first init. Follow wl.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agob43: N-PHY: determine usage of radio regulatory workaround correctly
Rafał Miłecki [Fri, 22 Oct 2010 15:43:48 +0000 (17:43 +0200)]
b43: N-PHY: determine usage of radio regulatory workaround correctly

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agob43: define known SPROM boardflags2 bits
Rafał Miłecki [Fri, 22 Oct 2010 15:43:47 +0000 (17:43 +0200)]
b43: define known SPROM boardflags2 bits

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agob43: N-PHY: fix 2055 radio init
Rafał Miłecki [Fri, 22 Oct 2010 15:43:46 +0000 (17:43 +0200)]
b43: N-PHY: fix 2055 radio init

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoiwlwifi: avoid commit rxon during scan in iwl_set_no_assoc
Stanislaw Gruszka [Fri, 22 Oct 2010 15:04:30 +0000 (17:04 +0200)]
iwlwifi: avoid commit rxon during scan in iwl_set_no_assoc

Currently we are canceling scan when changing BSSID. Behave the same
when changing association and beacon enablement, to avoid committing
rxon during scan in iwl_set_no_assoc().

Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Acked-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoiwlwifi: defer update power mode while scan
Stanislaw Gruszka [Fri, 22 Oct 2010 15:04:29 +0000 (17:04 +0200)]
iwlwifi: defer update power mode while scan

Do not set power mode when scanning, and defer that when scan finish.
We still set power mode in force case i.e. when device is overheated.

Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Acked-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoiwlwifi: avoid commit rxon during scan in iwlagn_bt_traffic_change_work
Stanislaw Gruszka [Fri, 22 Oct 2010 15:04:28 +0000 (17:04 +0200)]
iwlwifi: avoid commit rxon during scan in iwlagn_bt_traffic_change_work

Avoid sending commands to firmware (including commit_rxon) when scan
is pending and we are calling iwlagn_bt_traffic_change_work simultaneously.

Also comment some innocent race conditions.

Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Acked-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoiwlwifi: avoid commit rxon during scan in iwlagn_configure_filter
Stanislaw Gruszka [Fri, 22 Oct 2010 15:04:27 +0000 (17:04 +0200)]
iwlwifi: avoid commit rxon during scan in iwlagn_configure_filter

Almost anywhere in the code we avoid committing rxon while performing
scan, and make rxon commit when scan complete. However in some places
in the code we do not follow that rule. This patch fix that problem in
iwlagn_configure_filter().

Since we do not commit directly in iwl3945_configure_filter, we can
also do the same for agn, so I just remove iwlcore_commit_rxon()
function and add a comment. Also change comment for iwl3945.

Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Acked-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoiwlwifi: fix set_tx_power vs scan
Stanislaw Gruszka [Fri, 22 Oct 2010 15:04:26 +0000 (17:04 +0200)]
iwlwifi: fix set_tx_power vs scan

According to comment in iwl_bg_scan_completed, setting tx power should
be deferred during pending scan, but we are not doing this.

This patch change code to really defer setting tx power after scan
complete. Additionally refactor iwl_set_tx_power code and call
lib->send_tx_power() directly from iwlagn_commit_rxon.

Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Acked-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoiwlwifi: send tx_power_cmd synchronously
Stanislaw Gruszka [Fri, 22 Oct 2010 15:04:25 +0000 (17:04 +0200)]
iwlwifi: send tx_power_cmd synchronously

On 5xxx and 6xxx change to send tx_power_cmd command synchronously,
to do not start other commands when setting tx power is pending.
We currently do the same for 4956 and 3945.

Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Acked-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoiwlwifi: warn when send tx power settings during scan
Stanislaw Gruszka [Mon, 25 Oct 2010 08:34:50 +0000 (10:34 +0200)]
iwlwifi: warn when send tx power settings during scan

Add WARN_ONCE when scanning is pending. Use STATUS_SCAN_HW bit since we
can have scan canceled or completed but STATUS_SCANNING bit still set.

v1 -> v2: replace EIO to EAGAIN

Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Acked-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: minstrel_ht - reduce the overhead of rate sampling
Felix Fietkau [Thu, 21 Oct 2010 00:47:25 +0000 (02:47 +0200)]
mac80211: minstrel_ht - reduce the overhead of rate sampling

- reduce the number of retransmission attempts for sample rates
- sample lower rates less often
- do not use RTS/CTS for sampling frames
- increase the time between sampling attempts

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: fix potential spurious tx error bit interpretation
Felix Fietkau [Thu, 21 Oct 2010 00:47:23 +0000 (02:47 +0200)]
ath9k_hw: fix potential spurious tx error bit interpretation

According to documentation, AR_ExcessiveRetries, AR_Filtered and
AR_FIFOUnderrun are only valid if AR_FrmXmitOK is clear.

Not checking this might result in suboptimal FIFO settings, unnecessary
retransmissions, or other connectivity issues.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Cc: stable@kernel.org
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: add a debug warning when we cannot stop RX
Luis R. Rodriguez [Wed, 20 Oct 2010 23:07:08 +0000 (16:07 -0700)]
ath9k: add a debug warning when we cannot stop RX

We have seen several DMA races when we race against
stopping and starting the PCU. I suspect that when
we cannot stop the PCU we may hit some of these same
races so warn against them for now but only when
debugging (CONFIG_ATH_DEBUG) is enabled.

If you run into this warning and are a developer,
please fix the cause of the warning. The potential
here, although I cannot prove yet, is that the DMA
engine can be confused and start writing to a buffer
that was already DMA'd before and at least the kernel
assumes is not being accessed by hardware anymore.

Cc: Ben Greear <greearb@candelatech.com>
Cc: Kyungwan Nam <kyungwan.nam@atheros.com>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath: add a ATH_DBG_WARN()
Luis R. Rodriguez [Thu, 21 Oct 2010 19:10:11 +0000 (12:10 -0700)]
ath: add a ATH_DBG_WARN()

To be used to throw out warnings only for developers.
This can be used by some corner cases that developers
already know can be hit but developers want to address
so to avoid spewing out a warning this can only be
enabled with CONFIG_ATH_DEBUG enabled.

Cc: Ben Greear <greearb@candelatech.com>
Cc: Kyungwan Nam <kyungwan.nam@atheros.com>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agocfg80211: prefix REG_DBG_PRINT() with cfg80211
Luis R. Rodriguez [Wed, 20 Oct 2010 17:18:59 +0000 (10:18 -0700)]
cfg80211: prefix REG_DBG_PRINT() with cfg80211

Everyone's doing it, its the cool thing.

Cc: Easwar Krishnan <easwar.krishnan@atheros.com>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agocfg80211: add debug print when processing a channel
Luis R. Rodriguez [Thu, 21 Oct 2010 13:47:04 +0000 (19:17 +0530)]
cfg80211: add debug print when processing a channel

In the worst case you are seeing really odd things you want
more information than what is provided right now, for those
that insist and want debug info through CONFIG_CFG80211_REG_DEBUG
provide a print of when we are processing a channel and with what
regulatory rule.

Cc: Easwar Krishnan <easwar.krishnan@atheros.com>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: Senthil Balasubramanian <senthilkumar@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agocfg80211: add debug print when disabling a channel on a custom regd
Luis R. Rodriguez [Wed, 20 Oct 2010 17:18:57 +0000 (10:18 -0700)]
cfg80211: add debug print when disabling a channel on a custom regd

Cc: Easwar Krishnan <easwar.krishnan@atheros.com>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agocfg80211: add debug prints for when we ignore regulatory hints
Luis R. Rodriguez [Thu, 21 Oct 2010 13:47:03 +0000 (19:17 +0530)]
cfg80211: add debug prints for when we ignore regulatory hints

This can help with debugging issues. You will only see
these with CONFIG_CFG80211_REG_DEBUG enabled.

Cc: Easwar Krishnan <easwar.krishnan@atheros.com>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: Senthil Balasubramanian <senthilkumar@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agocfg80211: fix disabling channels based on hints
Luis R. Rodriguez [Wed, 20 Oct 2010 17:18:55 +0000 (10:18 -0700)]
cfg80211: fix disabling channels based on hints

After a module loads you will have loaded the world roaming regulatory
domain or a custom regulatory domain. Further regulatory hints are
welcomed and should be respected unless the regulatory hint is coming
from a country IE as the IEEE spec allows for a country IE to be a subset
of what is allowed by the local regulatory agencies.

So disable all channels that do not fit a regulatory domain sent
from a unless the hint is from a country IE and the country IE had
no information about the band we are currently processing.

This fixes a few regulatory issues, for example for drivers that depend
on CRDA and had no 5 GHz freqencies allowed were not properly disabling
5 GHz at all, furthermore it also allows users to restrict devices
further as was intended.

If you recieve a country IE upon association we will also disable the
channels that are not allowed if the country IE had at least one
channel on the respective band we are procesing.

This was the original intention behind this design but it was
completely overlooked...

Cc: David Quan <david.quan@atheros.com>
Cc: Jouni Malinen <jouni.malinen@atheros.com>
cc: Easwar Krishnan <easwar.krishnan@atheros.com>
Cc: stable@kernel.org
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agocfg80211: fix allowing country IEs for WIPHY_FLAG_STRICT_REGULATORY
Luis R. Rodriguez [Wed, 20 Oct 2010 17:18:54 +0000 (10:18 -0700)]
cfg80211: fix allowing country IEs for WIPHY_FLAG_STRICT_REGULATORY

We should be enabling country IE hints for WIPHY_FLAG_STRICT_REGULATORY
even if we haven't yet recieved regulatory domain hint for the driver
if it needed one. Without this Country IEs are not passed on to drivers
that have set WIPHY_FLAG_STRICT_REGULATORY, today this is just all
Atheros chipset drivers: ath5k, ath9k, ar9170, carl9170.

This was part of the original design, however it was completely
overlooked...

Cc: Easwar Krishnan <easwar.krishnan@atheros.com>
Cc: stable@kernel.org
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agocfg80211: pass the reg hint initiator to helpers
Luis R. Rodriguez [Wed, 20 Oct 2010 17:18:53 +0000 (10:18 -0700)]
cfg80211: pass the reg hint initiator to helpers

This is required later.

Cc: Easwar Krishnan <easwar.krishnan@atheros.com>
Cc: stable@kernel.org
signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>

Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agorfkill: remove dead code
Stephen Hemminger [Wed, 20 Oct 2010 17:16:58 +0000 (10:16 -0700)]
rfkill: remove dead code

The following code is defined but never used.

Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agossb: Clear RETRY_TIMEOUT in PCI Configuration for normal devices
Larry Finger [Wed, 20 Oct 2010 14:59:33 +0000 (09:59 -0500)]
ssb: Clear RETRY_TIMEOUT in PCI Configuration for normal devices

MMIO log traces obtained using the Broadcom wl hybrid driver show that
the RETRY_TIMEOUT register (0x41) in PCI configuration space is cleared
if non-zero. Similar code found in other drivers such as ipw2100 show
this operation is needed to keep PCI Tx retries from interfering with
C3 CPU state. There are no known cases where omission of this code has
caused a problem, but this patch is offered just in case such a situation
occurs.

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath5k: make ath5k_update_bssid_mask_and_opmode() static
Luis R. Rodriguez [Wed, 20 Oct 2010 13:59:38 +0000 (06:59 -0700)]
ath5k: make ath5k_update_bssid_mask_and_opmode() static

This fixes this sparse warning:

  CHECK   drivers/net/wireless/ath/ath5k/base.c
drivers/net/wireless/ath/ath5k/base.c:569:6: warning: symbol
'ath5k_update_bssid_mask_and_opmode' was not declared. Should it be static?

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath: make ath_hw_set_keycache_entry() static
Luis R. Rodriguez [Wed, 20 Oct 2010 13:59:37 +0000 (06:59 -0700)]
ath: make ath_hw_set_keycache_entry() static

This fixes this sparse warning:

  CHECK   drivers/net/wireless/ath/key.c
drivers/net/wireless/ath/key.c:110:6: warning: symbol
'ath_hw_set_keycache_entry' was not declared. Should it be static?

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath: make ath_hw_keysetmac() static
Luis R. Rodriguez [Wed, 20 Oct 2010 13:59:36 +0000 (06:59 -0700)]
ath: make ath_hw_keysetmac() static

This fixes this sparse warning:

  CHECK   drivers/net/wireless/ath/key.c
drivers/net/wireless/ath/key.c:70:6: warning: symbol 'ath_hw_keysetmac' was not declared. Should it be static?

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: initialize per-channel tx power limits instead of hardcoding them
Felix Fietkau [Wed, 20 Oct 2010 00:09:46 +0000 (02:09 +0200)]
ath9k: initialize per-channel tx power limits instead of hardcoding them

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: extend ath9k_hw_set_txpowerlimit to test channel txpower
Felix Fietkau [Wed, 20 Oct 2010 01:08:53 +0000 (03:08 +0200)]
ath9k_hw: extend ath9k_hw_set_txpowerlimit to test channel txpower

ath9k_hw_set_txpowerlimit gets an extra boolean parameter that - if set -
causes the rate txpower table and the regulatory limit to be calculated
and stored, without changing hardware registers.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: initialize regulatory->max_power_level in set_txpower for AR9003
Felix Fietkau [Wed, 20 Oct 2010 00:09:44 +0000 (02:09 +0200)]
ath9k_hw: initialize regulatory->max_power_level in set_txpower for AR9003

The same is done for the older chip families as well.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath5k: Add channel time to survey data
Bruno Randolf [Tue, 19 Oct 2010 07:56:54 +0000 (16:56 +0900)]
ath5k: Add channel time to survey data

Include the channel utilization (busy, rx, tx) in the survey results.

Signed-off-by: Bruno Randolf <br1@einfach.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath5k: Optimize descriptor alignment
Bruno Randolf [Tue, 19 Oct 2010 07:56:48 +0000 (16:56 +0900)]
ath5k: Optimize descriptor alignment

Similar to Felix Fietkau <nbd@openwrt.org> "ath9k_hw: optimize all descriptor
access functions" (13db2a80244908833502189a24de82a856668b8a).

Signed-off-by: Bruno Randolf <br1@einfach.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: Remove the median function in rate control
Mohammed Shafi Shajakhan [Mon, 18 Oct 2010 06:07:17 +0000 (11:37 +0530)]
ath9k: Remove the median function in rate control

With the current rate control selection method the median function is
nowhere used, so remove it.

Signed-off-by: Mohammed Shafi Shajakhan <mshajakhan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: optimize all descriptor access functions
Felix Fietkau [Fri, 15 Oct 2010 23:01:48 +0000 (01:01 +0200)]
ath9k_hw: optimize all descriptor access functions

Because all of the descriptor data structures are marked as __packed, GCC
assumes the worst case wrt. alignment and generates unaligned load/store
instructions on MIPS for access to all fields.
Since descriptors always have to be 4-byte-aligned, we can just mark the
data structures with __aligned(4), which allows GCC to generate much more
efficient code.
Verified through disassembly and OProfile comparisons.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: Convert to new PCI PM framework
Rafael J. Wysocki [Fri, 15 Oct 2010 22:36:17 +0000 (00:36 +0200)]
ath9k: Convert to new PCI PM framework

The ath9k driver uses the legacy PCI power management (suspend
and resume) callbacks that apparently cause intermittent problems
to happen (the adapter sometimes doesn't resume correctly on my
Acer Ferrari One).  Make it use the new PCI PM and let the PCI core
code handle the PCI-specific details of power transitions.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: fix sparse complaint on aphy for debugfs
Luis R. Rodriguez [Fri, 15 Oct 2010 20:27:49 +0000 (13:27 -0700)]
ath9k: fix sparse complaint on aphy for debugfs

This fixes this sparse complaint:

  CHECK   drivers/net/wireless/ath/ath9k/debug.c
drivers/net/wireless/ath/ath9k/debug.c:548:34: warning: symbol 'aphy' shadows an earlier one
drivers/net/wireless/ath/ath9k/debug.c:491:26: originally declared here

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: make ath9k_hw_gettsf32 static
Felix Fietkau [Fri, 15 Oct 2010 18:03:33 +0000 (20:03 +0200)]
ath9k_hw: make ath9k_hw_gettsf32 static

It is now only used in hw.c

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: remove a redundant call to ath9k_hw_gettsf32
Felix Fietkau [Fri, 15 Oct 2010 18:03:32 +0000 (20:03 +0200)]
ath9k: remove a redundant call to ath9k_hw_gettsf32

When the timer_next argument to ath9k_gen_timer_start is behind the tsf value,
tsf + timer_period is used, which is what ath_btcoex_period_timer was setting
it to.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: optimize/fix ANI RSSI processing
Felix Fietkau [Fri, 15 Oct 2010 18:03:31 +0000 (20:03 +0200)]
ath9k: optimize/fix ANI RSSI processing

ANI needs the RSSI average only in station mode, and only for tracking
the signal strength of beacons of the AP that it is connected to.
Adjust the code to track on the beacon RSSI, and store the average of that
in the ath_wiphy struct.
With these changes, we can get rid of this extra station lookup in the
rx path, which saves precious CPU cycles.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: optimize tx status descriptor processing
Felix Fietkau [Fri, 15 Oct 2010 18:03:30 +0000 (20:03 +0200)]
ath9k_hw: optimize tx status descriptor processing

Disassembly shows, that at least on MIPS, the compiler generates a lot of
memory accesses to the same location in the descriptor field parsing.
Since it is operating on uncached memory, this can be quite expensive in
this hot path.
Change the code a bit to help the compiler optimize it properly, and get
rid of some unused fields in the ath_tx_status struct.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: small optimization in ar9002_hw_get_isr
Felix Fietkau [Fri, 15 Oct 2010 18:03:29 +0000 (20:03 +0200)]
ath9k_hw: small optimization in ar9002_hw_get_isr

ah->config.rx_intr_mitigation does not need to be checked before checking
the rx interrupt mask for AR_ISR_RXMINTR or AR_ISR_RXINTM, as those
interrupts will be masked out if rx interrupt mitigation is disabled.

Avoid reading AR_ISR_S5_S twice by reordering the code to be more concise.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: optimize interrupt mask changes
Felix Fietkau [Mon, 8 Nov 2010 19:54:47 +0000 (20:54 +0100)]
ath9k_hw: optimize interrupt mask changes

OProfile showed that ath9k was spending way too much time in
ath9k_hw_set_interrupts. Since most of the interrupt mask changes only
need to globally enable/disable interrupts, it makes sense to split
this part into separate functions, replacing all calls to
ath9k_hw_set_interrupts(ah, 0) with ath9k_hw_disable_interrupts(ah).

ath9k_hw_set_interrupts(ah, ah->imask) only gets changed to
ath9k_hw_enable_interrupts(ah), whenever ah->imask was not changed
since the point where interrupts were disabled.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agob43: N-PHY: define registers names for 2056 radio
Rafał Miłecki [Thu, 14 Oct 2010 21:04:40 +0000 (23:04 +0200)]
b43: N-PHY: define registers names for 2056 radio

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Cc: Henry Ptasinski <henryp@broadcom.com>
Cc: Brett Rudley <brudley@broadcom.com>
Cc: Nohee Ko <noheek@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agob43: N-PHY: define registers names for 2056 radio
Rafał Miłecki [Sun, 17 Oct 2010 13:38:51 +0000 (15:38 +0200)]
b43: N-PHY: define registers names for 2056 radio

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Cc: Henry Ptasinski <henryp@broadcom.com>
Cc: Brett Rudley <brudley@broadcom.com>
Cc: Nohee Ko <noheek@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agob43: N-PHY: define channel table struct for rev3+ devices
Rafał Miłecki [Thu, 14 Oct 2010 21:01:02 +0000 (23:01 +0200)]
b43: N-PHY: define channel table struct for rev3+ devices

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoLinux 2.6.37-rc1 v2.6.37-rc1
Linus Torvalds [Mon, 1 Nov 2010 11:54:12 +0000 (07:54 -0400)]
Linux 2.6.37-rc1

13 years agoMerge branch 'for-linus' of git://www.jni.nu/cris
Linus Torvalds [Mon, 1 Nov 2010 11:53:19 +0000 (07:53 -0400)]
Merge branch 'for-linus' of git://www.jni.nu/cris

* 'for-linus' of git://www.jni.nu/cris:
  CRIS: Add paths for CRISv10 serial driver
  CRIS: Fix RS485 delay handling.
  Add missing "struct" to in sizeof.

13 years agoMerge branch 'next-spi' of git://git.secretlab.ca/git/linux-2.6
Linus Torvalds [Mon, 1 Nov 2010 11:50:43 +0000 (07:50 -0400)]
Merge branch 'next-spi' of git://git.secretlab.ca/git/linux-2.6

* 'next-spi' of git://git.secretlab.ca/git/linux-2.6:
  spi/pl022: fix erroneous platform data in U300
  spi: fixed odd static string conventions in core code
  spi/bfin_spi: only request GPIO on first load
  spi/bfin_spi: handle error/status changes after data interrupts
  spi: enable spi_board_info to be registered after spi_master

13 years agoCRIS: Add paths for CRISv10 serial driver
Jesper Nilsson [Mon, 1 Nov 2010 10:01:14 +0000 (11:01 +0100)]
CRIS: Add paths for CRISv10 serial driver

Signed-off-by: Jesper Nilsson <jesper.nilsson@axis.com>
13 years agospi/pl022: fix erroneous platform data in U300
Linus Walleij [Thu, 21 Oct 2010 19:05:25 +0000 (21:05 +0200)]
spi/pl022: fix erroneous platform data in U300

This fixes an erroneous use of LSB first in the U300 machine, the
PL022 used in U300 is a standard ARM core that doesn't support this
bit so it should never have been set.

Cc: Kevin Wells <wellsk40@gmail.com>OA
Signed-off-by: Linus Walleij <linus.walleij@stericsson.com>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
13 years agoInput: ir-keytable - fix uninitialized variable warning
Dmitry Torokhov [Sun, 31 Oct 2010 22:18:42 +0000 (15:18 -0700)]
Input: ir-keytable - fix uninitialized variable warning

We were forgetting to set up proper return value in success path causing
ir_getkeycode() to fail intermittently:

drivers/media/IR/ir-keytable.c: In function 'ir_getkeycode':
drivers/media/IR/ir-keytable.c:363: warning: 'retval' may be used
uninitialized in this function

Reported-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agoMerge branch 'i2c-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelvar...
Linus Torvalds [Mon, 1 Nov 2010 01:03:23 +0000 (21:03 -0400)]
Merge branch 'i2c-for-linus' of git://git./linux/kernel/git/jdelvare/staging

* 'i2c-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelvare/staging:
  i2c-i801: Add PCI idents for Patsburg 'IDF' SMBus controllers
  i2c-i801: Handle multiple instances instead of keeping global state
  i2c-i801: Add Intel Patsburg device ID
  i2c: Drop unused I2C_CLASS_TV flags

13 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input
Linus Torvalds [Mon, 1 Nov 2010 01:02:39 +0000 (21:02 -0400)]
Merge branch 'for-linus' of git://git./linux/kernel/git/dtor/input

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
  Input: appletouch - remove extra KERN_DEBUG use from dprintk
  Input: bu21013_ts - fix null dereference in error handling
  Input: ad7879 - prevent invalid finger data reports

13 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6
Linus Torvalds [Mon, 1 Nov 2010 00:43:12 +0000 (20:43 -0400)]
Merge git://git./linux/kernel/git/jejb/scsi-misc-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6: (70 commits)
  [SCSI] pmcraid: add support for set timestamp command and other fixes
  [SCSI] pmcraid: remove duplicate struct member
  [SCSI] qla4xxx: Fix cmd check in qla4xxx_cmd_wait
  [SCSI] megaraid_sas: Version and documentation update
  [SCSI] megaraid_sas: Add three times Online controller reset
  [SCSI] megaraid_sas: Add input parameter for max_sectors
  [SCSI] megaraid_sas: support devices update flag
  [SCSI] libosd: write/read_sg_kern API
  [SCSI] libosd: Support for scatter gather write/read commands
  [SCSI] libosd: Free resources in reverse order of allocation
  [SCSI] libosd: Fix bug in attr_page handling
  [SCSI] lpfc 8.3.18: Update lpfc driver version to 8.3.18
  [SCSI] lpfc 8.3.18: Add new WQE support
  [SCSI] lpfc 8.3.18: Fix critical errors
  [SCSI] lpfc 8.3.18: Adapter Shutdown and Unregistration cleanup
  [SCSI] lpfc 8.3.18: Add logic to detect last devloss timeout
  [SCSI] lpfc 8.3.18: Add support of received ELS commands
  [SCSI] lpfc 8.3.18: FC/FCoE Discovery fixes
  [SCSI] ipr: add definitions for a new adapter
  [SCSI] bfa: fix comments for c files
  ...

13 years agoMerge branches 'irq-core-for-linus' and 'core-locking-for-linus' of git://git.kernel...
Linus Torvalds [Mon, 1 Nov 2010 00:40:24 +0000 (20:40 -0400)]
Merge branches 'irq-core-for-linus' and 'core-locking-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'irq-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  genirq: Fix up irq_node() for irq_data changes.
  genirq: Add single IRQ reservation helper
  genirq: Warn if enable_irq is called before irq is set up

* 'core-locking-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  semaphore: Remove mutex emulation
  staging: Final semaphore cleanup
  jbd2: Convert jbd2_slab_create_sem to mutex
  hpfs: Convert sbi->hpfs_creation_de to mutex

Fix up trivial change/delete conflicts with deleted 'dream' drivers
(drivers/staging/dream/camera/{mt9d112.c,mt9p012_fox.c,mt9t013.c,s5k3e2fx.c})

13 years agoi2c-i801: Add PCI idents for Patsburg 'IDF' SMBus controllers
David Woodhouse [Sun, 31 Oct 2010 20:07:00 +0000 (21:07 +0100)]
i2c-i801: Add PCI idents for Patsburg 'IDF' SMBus controllers

These are the extra 'Integrated Device Function' SMBus controllers found
on the Patsburg chipset. Mention the absence of slave mode support.

Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
13 years agoi2c-i801: Handle multiple instances instead of keeping global state
David Woodhouse [Sun, 31 Oct 2010 20:06:59 +0000 (21:06 +0100)]
i2c-i801: Handle multiple instances instead of keeping global state

It's poor form to keep driver state in global variables rather than
per-instance. It never really mattered in practice when there was only
one controller on the chipset, but the latest chipsets do have more
than one controller, so now we care.

Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
13 years agoi2c-i801: Add Intel Patsburg device ID
Seth Heasley [Sun, 31 Oct 2010 20:06:59 +0000 (21:06 +0100)]
i2c-i801: Add Intel Patsburg device ID

Add support for the Intel Patsburg PCH SMBus Controller.

Signed-off-by: Seth Heasley <seth.heasley@intel.com>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
13 years agoi2c: Drop unused I2C_CLASS_TV flags
Jean Delvare [Sun, 31 Oct 2010 20:06:59 +0000 (21:06 +0100)]
i2c: Drop unused I2C_CLASS_TV flags

There are no users left for I2C_CLASS_TV_ANALOG and
I2C_CLASS_TV_DIGITAL, so we can get rid of them.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
13 years agoInput: appletouch - remove extra KERN_DEBUG use from dprintk
Joe Perches [Sun, 31 Oct 2010 00:19:49 +0000 (17:19 -0700)]
Input: appletouch - remove extra KERN_DEBUG use from dprintk

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
13 years agolocks: remove fl_copy_lock lock_manager operation
Christoph Hellwig [Sun, 31 Oct 2010 12:35:31 +0000 (08:35 -0400)]
locks: remove fl_copy_lock lock_manager operation

This one was only used for a nasty hack in nfsd, which has recently
been removed.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agolocks: let the caller free file_lock on ->setlease failure
Christoph Hellwig [Sun, 31 Oct 2010 12:35:10 +0000 (08:35 -0400)]
locks: let the caller free file_lock on ->setlease failure

The caller allocated it, the caller should free it.

The only issue so far is that we could change the flp pointer even on an
error return if the fl_change callback failed.  But we can simply move
the flp assignment after the fl_change invocation, as the callers don't
care about the flp return value if the setlease call failed.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agoinitramfs: Fix initramfs size for 32-bit arches
Geert Uytterhoeven [Sun, 31 Oct 2010 09:56:23 +0000 (10:56 +0100)]
initramfs: Fix initramfs size for 32-bit arches

Commit ffe8018c3424 ("initramfs: fix initramfs size calculation") broke
32-bit big-endian arches like (on ARAnyM):

    VFS: Cannot open root device "hda1" or unknown-block(3,1)
    Please append a correct "root=" boot option; here are the available partitions:
    fe80         1059408 nfhd8  (driver?)
      fe81          921600 nfhd8p1 00000000-0000-0000-0000-000000000nfhd8p1
      fe82          137807 nfhd8p2 00000000-0000-0000-0000-000000000nfhd8p2
    0200            3280 fd0  (driver?)
    0201            3280 fd1  (driver?)
    0300         1059408 hda  driver: ide-gd
      0301          921600 hda1 00000000-0000-0000-0000-000000000hda1
      0302          137807 hda2 00000000-0000-0000-0000-000000000hda2
    Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(3,1)

As pointed out by Kerstin Jonsson <kerstin.jonsson@ericsson.com>, this
is due to CONFIG_32BIT not being defined, so the initramfs size field is
done as a 64-bit quad.  On little-endian (like x86) this doesn matter,
but on a big-endian machine the 32-bit reads will see the (zero) high
bits.

Only mips, s390, and score set CONFIG_32BIT for 32-bit builds, so fix it for
all other 32-bit arches by inverting the logic and testing for CONFIG_64BIT,
which should be defined on all 64-bit arches.

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
[ I think we should just make it "u64" on all architectures and get
  rid of the whole #ifdef CONFIG_xxBIT   - Linus ]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Sun, 31 Oct 2010 01:42:58 +0000 (18:42 -0700)]
Merge git://git./linux/kernel/git/davem/net-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6:
  isdn: mISDN: socket: fix information leak to userland
  netdev: can: Change mail address of Hans J. Koch
  pcnet_cs: add new_id
  net: Truncate recvfrom and sendto length to INT_MAX.
  RDS: Let rds_message_alloc_sgs() return NULL
  RDS: Copy rds_iovecs into kernel memory instead of rereading from userspace
  RDS: Clean up error handling in rds_cmsg_rdma_args
  RDS: Return -EINVAL if rds_rdma_pages returns an error
  net: fix rds_iovec page count overflow
  can: pch_can: fix section mismatch warning by using a whitelisted name
  can: pch_can: fix sparse warning
  netxen_nic: Fix the tx queue manipulation bug in netxen_nic_probe
  ip_gre: fix fallback tunnel setup
  vmxnet: trivial annotation of protocol constant
  vmxnet3: remove unnecessary byteswapping in BAR writing macros
  ipv6/udp: report SndbufErrors and RcvbufErrors
  phy/marvell: rename 88ec048 to 88e1318s and fix mscr1 addr

13 years agonfsd4: initialize delegation pointer to lease
J. Bruce Fields [Sat, 30 Oct 2010 21:31:16 +0000 (17:31 -0400)]
nfsd4: initialize delegation pointer to lease

The NFSv4 server was initializing the dp->dl_flock pointer by the
somewhat ridiculous method of a locks_copy_lock callback.

Now that setlease uses the passed-in lock instead of doing a copy,
dl_flock no longer gets set, resulting in the lock leaking on delegation
release, and later possible hangs (among other problems).

So, initialize dl_flock and get rid of the callback.

Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agolocks: fix setlease methods to free passed-in lock
J. Bruce Fields [Sat, 30 Oct 2010 21:31:15 +0000 (17:31 -0400)]
locks: fix setlease methods to free passed-in lock

We modified setlease to require the caller to allocate the new lease in
the case of creating a new lease, but forgot to fix up the filesystem
methods.

Cc: Steven Whitehouse <swhiteho@redhat.com>
Cc: Steve French <sfrench@samba.org>
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agolocks: fix leaks on setlease errors
J. Bruce Fields [Sat, 30 Oct 2010 21:31:14 +0000 (17:31 -0400)]
locks: fix leaks on setlease errors

We're depending on setlease to free the passed-in lease on failure.

Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agolocks: prevent ENOMEM on lease unlock
J. Bruce Fields [Sat, 30 Oct 2010 21:31:13 +0000 (17:31 -0400)]
locks: prevent ENOMEM on lease unlock

Removing a lock shouldn't require any allocations; a failure due to
ENOMEM leaves the caller with a choice between retrying or giving up and
leaking an unused lease.

Next we should split the other lease calls into add and delete cases.
I wanted to start with just the bugfix.

Signed-off-by: J. Bruce Fields <bfields@redhat.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agodrivers/media/IR/ir-keytable.c: fix binary search
David Härdeman [Sat, 30 Oct 2010 20:17:44 +0000 (22:17 +0200)]
drivers/media/IR/ir-keytable.c: fix binary search

The input-large-scancode patches changed the binary search in
drivers/media/IR/ir-keytable.c to use unsigned integers, but
signed integers are actually necessary for the algorithm to work.

Signed-off-by: David Härdeman <david@hardeman.nu>
Cc: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agoisdn: mISDN: socket: fix information leak to userland
Kulikov Vasiliy [Fri, 29 Oct 2010 23:04:33 +0000 (23:04 +0000)]
isdn: mISDN: socket: fix information leak to userland

Structure mISDN_devinfo is copied to userland with the field "name"
that has the last elements unitialized.  It leads to leaking of
contents of kernel stack memory.

Signed-off-by: Vasiliy Kulikov <segooon@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agonetdev: can: Change mail address of Hans J. Koch
Hans J. Koch [Fri, 29 Oct 2010 12:33:57 +0000 (12:33 +0000)]
netdev: can: Change mail address of Hans J. Koch

My old mail address doesn't exist anymore. This changes all occurrences
to my new address.

Signed-off-by: Hans J. Koch <hjk@hansjkoch.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agopcnet_cs: add new_id
Ken Kawasaki [Fri, 29 Oct 2010 12:17:51 +0000 (12:17 +0000)]
pcnet_cs: add new_id

pcnet_cs:
    add new_id: "corega Ether CF-TD" 10Base-T PCMCIA card.

Signed-off-by: Ken Kawasaki <ken_kawasaki@spring.nifty.jp>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agonet: Truncate recvfrom and sendto length to INT_MAX.
Linus Torvalds [Sat, 30 Oct 2010 23:43:10 +0000 (16:43 -0700)]
net: Truncate recvfrom and sendto length to INT_MAX.

Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoRDS: Let rds_message_alloc_sgs() return NULL
Andy Grover [Thu, 28 Oct 2010 15:40:59 +0000 (15:40 +0000)]
RDS: Let rds_message_alloc_sgs() return NULL

Even with the previous fix, we still are reading the iovecs once
to determine SGs needed, and then again later on. Preallocating
space for sg lists as part of rds_message seemed like a good idea
but it might be better to not do this. While working to redo that
code, this patch attempts to protect against userspace rewriting
the rds_iovec array between the first and second accesses.

The consequences of this would be either a too-small or too-large
sg list array. Too large is not an issue. This patch changes all
callers of message_alloc_sgs to handle running out of preallocated
sgs, and fail gracefully.

Signed-off-by: Andy Grover <andy.grover@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoRDS: Copy rds_iovecs into kernel memory instead of rereading from userspace
Andy Grover [Thu, 28 Oct 2010 15:40:58 +0000 (15:40 +0000)]
RDS: Copy rds_iovecs into kernel memory instead of rereading from userspace

Change rds_rdma_pages to take a passed-in rds_iovec array instead
of doing copy_from_user itself.

Change rds_cmsg_rdma_args to copy rds_iovec array once only. This
eliminates the possibility of userspace changing it after our
sanity checks.

Implement stack-based storage for small numbers of iovecs, based
on net/socket.c, to save an alloc in the extremely common case.

Although this patch reduces iovec copies in cmsg_rdma_args to 1,
we still do another one in rds_rdma_extra_size. Getting rid of
that one will be trickier, so it'll be a separate patch.

Signed-off-by: Andy Grover <andy.grover@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoRDS: Clean up error handling in rds_cmsg_rdma_args
Andy Grover [Thu, 28 Oct 2010 15:40:57 +0000 (15:40 +0000)]
RDS: Clean up error handling in rds_cmsg_rdma_args

We don't need to set ret = 0 at the end -- it's initialized to 0.

Also, don't increment s_send_rdma stat if we're exiting with an
error.

Signed-off-by: Andy Grover <andy.grover@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoRDS: Return -EINVAL if rds_rdma_pages returns an error
Andy Grover [Thu, 28 Oct 2010 15:40:56 +0000 (15:40 +0000)]
RDS: Return -EINVAL if rds_rdma_pages returns an error

rds_cmsg_rdma_args would still return success even if rds_rdma_pages
returned an error (or overflowed).

Signed-off-by: Andy Grover <andy.grover@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agonet: fix rds_iovec page count overflow
Linus Torvalds [Thu, 28 Oct 2010 15:40:55 +0000 (15:40 +0000)]
net: fix rds_iovec page count overflow

As reported by Thomas Pollet, the rdma page counting can overflow.  We
get the rdma sizes in 64-bit unsigned entities, but then limit it to
UINT_MAX bytes and shift them down to pages (so with a possible "+1" for
an unaligned address).

So each individual page count fits comfortably in an 'unsigned int' (not
even close to overflowing into signed), but as they are added up, they
might end up resulting in a signed return value. Which would be wrong.

Catch the case of tot_pages turning negative, and return the appropriate
error code.

Reported-by: Thomas Pollet <thomas.pollet@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Andy Grover <andy.grover@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agocan: pch_can: fix section mismatch warning by using a whitelisted name
Marc Kleine-Budde [Sat, 30 Oct 2010 23:28:16 +0000 (16:28 -0700)]
can: pch_can: fix section mismatch warning by using a whitelisted name

This patch fixes the following section mismatch warning:

WARNING: drivers/net/can/pch_can.o(.data+0x18):
Section mismatch in reference from the variable pch_can_pcidev
to the variable .devinit.rodata:pch_pci_tbl
The variable pch_can_pcidev references
the variable __devinitconst pch_pci_tbl

This is actually a false positive which is fixed by giving the offending
variable a whitelisted name, it's renamed to "pch_can_pci_driver".
This makes sense because the variable is of the type "struct pci_driver".

Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Acked-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: David S. Miller <davem@davemloft.net>