* 'imx/driver' of git://git.linaro.org/people/shawnguo/linux-2.6: (3 commits)
ARM: imx6q: add cko1 clock
ARM: mxc: make imx_dma_is_general_purpose more generic for sdma
ARM: imx6: Rename DEBUG_IMX6Q_UART to UART4
This ensures the user won't encounter lag associated with getting in and
out of PSM when the card is in use.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Flush our Tx queues before suspending or changing the channel due to a
channel_switch element in the AP beacon.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Ensure our queues are empty at the end of a tx_flush(), in case we
timeout on passively waiting for them. This makes sure no left-overs are
transmitted when we are on the wrong channel.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Re-factor the Tx scheduler so that the system_hlid is taken into account
before restarting an iteration over the wlvifs. Previously this
hlid had a lower priority and would starve if some wlvif had many
packets.
In addition avoid iterating over wlvifs past last_wlvif when performing
the a second pass. If we had packets in those wlvifs they would have
been found earlier.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Before, the link was first freed (invalidating it in the map), and later
on vif removal, all valid wlvif-related links were reset. Since these
links were already invalid, we failed to reset them.
The bug was made worse by op_stop, which set the tx_queue_count to 0
arbitrarily. This resulted in a negative tx_queue_count in some scenarios.
Fix this by resetting the Tx-queues of a link when freeing it. Add a
WARN_ON and reset all link Tx-queues in op_stop, to avoid a negative
tx_queue_count.
[changed WARN_ON to WARN_ON_ONCE -- Luca]
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
These bits are used in op_tx to determine if a packet should be dropped.
As such we should use the spin lock to sync the state.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
With PSM handled in FW, the checks in wl1271_elp_work() are always true.
Thus during active traffic we constantly enter and exit ELP (many times
per second). As each ELP exit takes ~10ms, this can have an adverse
effect on throughput and interactivity.
Set the ELP timeout to the dyn-ps timeout. This period is longer and
avoids the above problem. It also makes sense to stay out of ELP while
we are awake on the network, to minimize delays in Tx/Rx. The same thing
was done by the mac80211 dynamic-ps mechanism before the FW DPS changes.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
This patch adds a new position_fix option value, 4, as a combo mode
to use LPIB for playbacks and POSBUF for captures. It's the way
recommended by Intel hardware guys.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
eMMa-PrP device included in Freescale i.MX2 chips can also
be used separately to process memory buffers. This patch
provides arch glue code for the driver which provides this
functionality.
Acked-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Javier Martin <javier.martin@vista-silicon.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
i.MX2x SoCs have a PrP which is capable of resizing and format
conversion of video frames. This driver provides support for
resizing and format conversion from YUYV to YUV420.
This operation is of the utmost importance since some of these
SoCs like i.MX27 include an H.264 video codec which only
accepts YUV420 as input.
Reviewed-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: Javier Martin <javier.martin@vista-silicon.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
This makes mount take slightly longer, but at the same time, the first
write to the filesystem will be faster too. It also means that if there
is a problem in the resource index, then we can refuse to mount rather
than having to try and report that when the first write occurs.
In addition, to avoid recursive locking, we hvae to take account of
instances when the rindex glock may already be held when we are
trying to update the rbtree of resource groups.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
This patch fixes a problem whereby gfs2_grow was failing and causing GFS2
to assert. The problem was that when GFS2's fallocate operation tried to
acquire an "allocation" it made sure the rindex was up to date, and if not,
it called gfs2_rindex_update. However, if the file being fallocated was
the rindex itself, it was already locked at that point. By calling
gfs2_rindex_update at an earlier point in time, we bring rindex up to date
and thereby avoid trying to lock it when the "allocation" is acquired.
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
This patch fixes a problem whereby you were unable to delete
files until other file system operations were done (such as
statfs, touch, writes, etc.) that caused the rindex to be
read in.
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
This patch fixes a narrow race window between the glock ref count
hitting zero and glocks being removed from the lru_list.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Thaw and poweroff routines are missing for spear-keyboard. They are
required for:
- Error case scenarios during freeze
- Using test features, of hibernate.
Signed-off-by: Viresh Kumar <viresh.kumar@st.com>
Signed-off-by: Rajeev Kumar <rajeev-dlh.kumar@st.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
Let platform pass mode information to keyboard driver according to which
it configures itself. The mode can be
- KEYPAD_9x9 0
- KEYPAD_6x6 1
- KEYPAD_2x2 2
Signed-off-by: Rajeev Kumar <rajeev-dlh.kumar@st.com>
Signed-off-by: Viresh Kumar <viresh.kumar@st.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
The major features of this series are:
- making RCU more aggressive about entering dyntick-idle mode in order to
improve energy efficiency
- converting a few more call_rcu()s to kfree_rcu()s
- applying a number of rcutree fixes and cleanups to rcutiny
- removing CONFIG_SMP #ifdefs from treercu
- allowing RCU CPU stall times to be set via sysfs
- adding CPU-stall capability to rcutorture
- adding more RCU-abuse diagnostics
- updating documentation
- fixing yet more issues located by the still-ongoing top-to-bottom
inspection of RCU, this time with a special focus on the
CPU-hotplug code path.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Register driver's ht capabilities in wiphy to avoid failures
in setting ht channels from hostapd.
Signed-off-by: Vasanthakumar Thiagarajan <vthiagar@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
Without this change, SDIO shuts down due to CRC error during
data communication to the firmware in some of the platform.
for example, scan request issued to the firmware doesn't
return scan completed events and cause the socket interface
to always return -16 (device busy).
SDIO pad drive strength should be reduced for hw2.1.1 board
to avoid such errors.
Signed-off-by: Raja Mani <rmani@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
This makes the device get into scan state as soon as a
scan request is received, instead of blocking the scan
request till the traffic comes down. It is necessary
for the deterministic foreground scan particularly when
having multiple vif operating where, if the scan is non-
deterministic, scan on one interface will not start
as long as there are traffic on the other interface. This
change passes 50 msec as foreground scan interval to fix this.
Signed-off-by: Vasanthakumar Thiagarajan <vthiagar@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
* 'drm-intel-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes/drm-intel:
drm/i915: Remove use of the autoreported ringbuffer HEAD position
drm/i915: Prevent a machine hang by checking crtc->active before loading lut
drm/i915: fix operator precedence when enabling RC6p
drm/i915: fix a sprite watermark computation to avoid divide by zero if xpos<0
drm/i915: fix mode set on load pipe. (v2)
htc_packet and htc_packet->buf_start are separately allocated
for endpoint 0. This is different for other endpoints where
packets are allocated as skb where htc_packet is skb->head
and they are freed properly. Free htc_packet and htc_packet->buf_start
separatly for endpoint 0.
Signed-off-by: Vasanthakumar Thiagarajan <vthiagar@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
Rx buffers should be allocated for control and best effort endpoints only
after the enpoints connection is esablished. But this is done before the
endpoint connection is complete, we don't even the control and BE endpoints
that time. Move the buffer allocation after endpoint connection is over,
after ath6kl_init_hw_start(). Found in review, never seen any real issue
with this.
Signed-off-by: Vasanthakumar Thiagarajan <vthiagar@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
We shouldn't hold any locks on release path. Pass a flag to
vhost_dev_cleanup to use the lockdep info correctly.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Sasha Levin <levinsasha928@gmail.com>
This is a tiny, but important, patch to vhost.
Vhost's worker thread only called schedule() when it had no work to do, and
it wanted to go to sleep. But if there's always work to do, e.g., the guest
is running a network-intensive program like netperf with small message sizes,
schedule() was *never* called. This had several negative implications (on
non-preemptive kernels):
1. Passing time was not properly accounted to the "vhost" process (ps and
top would wrongly show it using zero CPU time).
2. Sometimes error messages about RCU timeouts would be printed, if the
core running the vhost thread didn't schedule() for a very long time.
3. Worst of all, a vhost thread would "hog" the core. If several vhost
threads need to share the same core, typically one would get most of the
CPU time (and its associated guest most of the performance), while the
others hardly get any work done.
The trivial solution is to add
if (need_resched())
schedule();
After doing every piece of work. This will not do the heavy schedule() all
the time, just when the timer interrupt decided a reschedule is warranted
(so need_resched returns true).
Thanks to Abel Gordon for this patch.
Signed-off-by: Nadav Har'El <nyh@il.ibm.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Enables port E of IDT 92HD91 codec as output and sets correct output
phase between ports E and D and high pass filter.
Signed-off-by: Vitaliy Kulikov <Vitaliy.Kulikov@idt.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
gpio-pxa driver is shared among arch-pxa and arch-mmp. Clock is the
essential component on pxa3xx/pxa95x and arch-mmp. So we need to
define dummy clock in pxa25x/pxa27x instead.
Signed-off-by: Haojian Zhuang <haojian.zhuang@marvell.com>
IRQ name is required in gpio-pxa driver. So we need to append the name
in gpio device. Otherwise, we can't register gpio-pxa driver
successfully.
Signed-off-by: Haojian Zhuang <haojian.zhuang@marvell.com>
The hx4700 platform has 72 board gpios: 64 ASIC3 gpios numbered 192..255, and
8 EGPIO gpios numbered 256..263 (plus a 9th which is not used).
Thus the new CONFIG_ARCH_NR_GPIO config option must be set to 264.
Signed-off-by: Paul Parsons <lost.distance@yahoo.com>
Acked-by: Philipp Zabel <philipp.zabel@gmail.com>
Signed-off-by: Haojian Zhuang <haojian.zhuang@gmail.com>
Add LED support for the HTC ASIC3. Underlying support is provided by the mfd/asic3 and leds/leds-asic3 drivers. An example configuration is provided by the pxa/hx4700 platform.
Signed-off-by: Paul Parsons <lost.distance@yahoo.com>
Acked-by: Philipp Zabel <philipp.zabel@gmail.com>
Signed-off-by: Haojian Zhuang <haojian.zhuang@gmail.com>
The HP iPAQ hx4700 has 128Mb of flash provided by two Intel StrataFlash devices.
The hx4700 platform configuration defines a single 128Mb flash resource,
resulting in the MTD physmap-flash driver probing the first device only and
presuming the second device is identical:
physmap platform flash device: 08000000 at 00000000
physmap-flash: Found 2 x16 devices at 0x0 in 32-bit bank. Manufacturer ID 0x000089 Chip ID 0x008816
physmap-flash: Found 2 x16 devices at 0x4000000 in 32-bit bank
<snip>
erase region 0: offset=0x0,size=0x10000,blocks=4
erase region 1: offset=0x40000,size=0x40000,blocks=255
erase region 2: offset=0x4000000,size=0x10000,blocks=4
erase region 3: offset=0x4040000,size=0x40000,blocks=255
physmap-flash: 2 set(s) of 2 interleaved chips --> 32 partitions of 4096 KiB
Unfortunately the two devices are not identical. The first has a device ID of
0x8816, identifying a bottom parameter device. The second has a device ID of
0x8813, identifying a top parameter device. By not probing the second device,
physmap-flash does not discover the correct block sizes.
This patch splits the configuration into two 64Mb flash resources, forcing
physmap-flash to probe both devices and thus discover the correct block sizes:
physmap platform flash device: 04000000 at 00000000
physmap-flash: Found 2 x16 devices at 0x0 in 32-bit bank. Manufacturer ID 0x000089 Chip ID 0x008816
<snip>
erase region 0: offset=0x0,size=0x10000,blocks=4
erase region 1: offset=0x40000,size=0x40000,blocks=255
physmap-flash: 1 set(s) of 2 interleaved chips --> 16 partitions of 4096 KiB
physmap platform flash device: 04000000 at 04000000
physmap-flash: Found 2 x16 devices at 0x0 in 32-bit bank. Manufacturer ID 0x000089 Chip ID 0x008813
<snip>
erase region 0: offset=0x0,size=0x40000,blocks=255
erase region 1: offset=0x3fc0000,size=0x10000,blocks=4
physmap-flash: 1 set(s) of 2 interleaved chips --> 16 partitions of 4096 KiB
Concatenating MTD devices:
(0): "physmap-flash"
(1): "physmap-flash"
into device "physmap-flash"
Signed-off-by: Paul Parsons <lost.distance@yahoo.com>
Signed-off-by: Haojian Zhuang <haojian.zhuang@gmail.com>
sa1100-rtc driver could be shared among sa1100/pxa/mmp series silicon.
Since clk is used in mmp series silicon, add dummy clock support in
pxa also.
Signed-off-by: Haojian Zhuang <haojian.zhuang@marvell.com>
Avoid to hard coded irq in rtc-sa1100 driver since we could share
it among arch-sa1100/arch-pxa/arch-mmp.
We still keep hard coded register address since the requirement is
enabling both rtc-sa1100 and rtc-pxa driver. The register addresses
are conflict since they're only two wrappers on the same rtc device.
Signed-off-by: Haojian Zhuang <haojian.zhuang@marvell.com>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Since next alarm time is already calculated in rtc common interface, we
needn't keep this logic in rtc-sa1100.c any more.
Signed-off-by: Haojian Zhuang <haojian.zhuang@marvell.com>
Arnd Bergmann