This reverts commit d80c0d1418
(excluding the change to MAX_BL_NUM, which has since been removed).
The qcserial driver now enables the SET_CONTROL_LINE_STATE request
so that AT URCs will work. Move these devices back to the qcserial
driver, which is used for other devices in this series that follow
the same layout.
Cc: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David Ward <david.ward@ll.mit.edu>
Signed-off-by: Johan Hovold <johan@kernel.org>
This reverts commit 653cdc13a3.
The qcserial driver now enables the SET_CONTROL_LINE_STATE request
so that AT URCs will work. Move these devices back to the qcserial
driver, which is used for other devices in this series that follow
the same layout.
Cc: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David Ward <david.ward@ll.mit.edu>
Signed-off-by: Johan Hovold <johan@kernel.org>
Three Sierra Wireless modems have been found to require the CDC ACM
SET_CONTROL_LINE_STATE request on the AT port in order to receive
unsolicited response codes, most recently the Dell Wireless 5808e
(which is a re-branded Sierra Wireless EM7355). On the other hand,
the Sierra Wireless MC7710 does not seem to need this request, but
it was found to work either way.
Use this request on the AT port of devices with the Sierra Wireless
layout in the qcserial driver. The other modems that were moved to
the option driver to work around this can now be moved back.
Cc: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David Ward <david.ward@ll.mit.edu>
Signed-off-by: Johan Hovold <johan@kernel.org>
Only the option driver implements the send_setup callback; it uses the
SET_CONTROL_LINE_STATE request in CDC ACM to generate DTR/RTS signals
on the port. This is not driver-specific though and is needed by other
drivers, so move the function to the usb_wwan driver (with formatting
tweaks), and replace the callback pointer with a flag that enables the
request.
Suggested-by: Bjørn Mork <bjorn@mork.no>
Suggested-by: Johan Hovold <johan@kernel.org>
Signed-off-by: David Ward <david.ward@ll.mit.edu>
Signed-off-by: Johan Hovold <johan@kernel.org>
This is a partial, context modified revert of commit e463c6dda8
("USB: option: handle send_setup blacklisting at probe"), which
introduced an unnecessary struct option_private for storing the
interface number used in option_send_setup. Removing this struct
will allow option_send_setup to be generalized for other drivers.
Suggested-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David Ward <david.ward@ll.mit.edu>
Signed-off-by: Johan Hovold <johan@kernel.org>
Remove extra blank lines in the several places where functions were
separated by more than one.
Signed-off-by: Peter E. Berger <pberger@brimson.com>
Signed-off-by: Johan Hovold <johan@kernel.org>
Fix non-standard formatting in some of the comments.
Signed-off-by: Peter E. Berger <pberger@brimson.com>
[johan: minor fixes ]
Signed-off-by: Johan Hovold <johan@kernel.org>
This ensures that the link is not requesting any clock and the
PLL can turn off. The link is powered when controller is brought
out of reset.
Signed-off-by: Jeeja KP <jeeja.kp@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
On runtime pm resume, we need to download the firmware, also on
suspend we need to ensure all the interrupts from controller and
DSP are disabled.
Also since we download the firmware on resume, we don't need to do
so on init, so remove that bit
Signed-off-by: Jeeja KP <jeeja.kp@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Like we have in legacy mode HDA driver, we need to check the
status bit and handle interrupt only when it is not zero or all
bits set. We typically see the status as all 1's when controller
resumes from suspend, So add the check here as well and don't
handle for these cases.
Signed-off-by: Jeeja KP <jeeja.kp@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Skylake driver will set the SPA bit to 0 to turn off the DSP core.
Driver will poll the Current Power Active (CPA) bit to match the
Set Power Active (SPA) bit value. When CPA bit matches the value
of SPA bit, the achieved power state has reached.
In case of DSP power down, register that was polled is SPA
instead of CPA. This patch corrects the register to be polled
in case of DSP power down.
Signed-off-by: Jeeja KP <jeeja.kp@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Move request_firmware from edge_startup to download_fw.
Signed-off-by: Peter E. Berger <pberger@brimson.com>
Signed-off-by: Johan Hovold <johan@kernel.org>
Add support to create thermal zones based on the temperature sensors
provided by the SCP. The thermal zones can be defined using the
thermal DT bindings and should refer to the SCP sensor id to select
the sensor.
Signed-off-by: Punit Agrawal <punit.agrawal@arm.com>
Acked-by: Guenter Roeck <linux@roeck-us.net>
Cc: Eduardo Valentin <edubezval@gmail.com>
Create a driver to add support for SoC sensors exported by the System
Control Processor (SCP) via the System Control and Power Interface
(SCPI). The supported sensor types is one of voltage, temperature,
current, and power.
The sensor labels and values provided by the SCP are exported via the
hwmon sysfs interface.
Signed-off-by: Punit Agrawal <punit.agrawal@arm.com>
Acked-by: Guenter Roeck <linux@roeck-us.net>
Cc: Sudeep Holla <sudeep.holla@arm.com>
ARM System Control Processor (SCP) provides an API to query and use
the sensors available in the system. Extend the SCPI driver to support
sensor messages.
Signed-off-by: Punit Agrawal <punit.agrawal@arm.com>
Acked-by: Sudeep Holla <sudeep.holla@arm.com>
Separate the download and boot mode code from download_fw() into two new
helper functions: do_download_mode() and do_boot_mode().
Signed-off-by: Peter E. Berger <pberger@brimson.com>
Signed-off-by: Johan Hovold <johan@kernel.org>
Use serial->interface instead of serial->dev for messages in download_fw().
Signed-off-by: Peter E. Berger <pberger@brimson.com>
Signed-off-by: Johan Hovold <johan@kernel.org>
Remove unused "dev" parameter from build_i2c_fw_hdr() and its caller.
Signed-off-by: Peter E. Berger <pberger@brimson.com>
Signed-off-by: Johan Hovold <johan@kernel.org>
The System Control Processor (SCP) provides access to SoC sensors via
the System Control and Power Interface (SCPI) Message Protocol. Add
bindings to allow probing of these sensors. Also support referencing
of the sensors for setting up thermal zones via the thermal DT
bindings.
Signed-off-by: Punit Agrawal <punit.agrawal@arm.com>
Acked-by: Rob Herring <robh+dt@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Sudeep Holla <sudeep.holla@arm.com>
The SCP firmware on Juno provides access to SoC sensors via the
SCPI. Add the sensor nodes to the device tree to enable this support.
Signed-off-by: Punit Agrawal <punit.agrawal@arm.com>
Cc: Rob Herring <robh+dt@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Acked-by: Sudeep Holla <sudeep.holla@arm.com>
Acked-by: Liviu Dudau <liviu.dudau@arm.com>
This patch adds the CPU clocks so that the CPU DVFS can be enabled.
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
Acked-by: Liviu Dudau <Liviu.Dudau@arm.com>
Cc: Jon Medhurst (Tixy) <tixy@linaro.org>
This patch adds CPU topology on Juno. It will be useful for ther other
IP blocks depending on this topology.
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
Acked-by: Liviu Dudau <Liviu.Dudau@arm.com>
Cc: Jon Medhurst (Tixy) <tixy@linaro.org>
This patch adds support for the MHU mailbox peripheral used on Juno by
application processors to communicate with remote SCP handling most of
the CPU/system power management. It also adds the SRAM reserving the
shared memory and SCPI message protocol using that shared memory.
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
Acked-by: Liviu Dudau <Liviu.Dudau@arm.com>
Cc: Jon Medhurst (Tixy) <tixy@linaro.org>
WARN_ON() already contain an unlikely compiler flag. Drop it.
Signed-off-by: Geliang Tang <geliangtang@163.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
The DFS pattern detector ought to reset the
detector lines when a pulse is added with
lower time stamp than the previous (which
indicates a TSF restart).
This did not work so far and is fixed with
this patch.
The modification does not change detection
performance within the driver, since it
only ensures early reset (which is later
performed by the PRI detectors anyway).
It is relevant for synthetic tests and
statistical evaluations, where millions
of pulse patterns are processed and an
early reset helps reducing load.
Signed-off-by: Zefir Kurtisi <zefir.kurtisi@neratec.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
The check for send_pkt being NULL is redundant before the call
to htc_reclaim_txctrl_buf, therefore it should be removed. This was
detected by static analysis by cppcheck.
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
It's been reported that the atomic watermark series triggers some
regressions on SKL, which we haven't been able to track down yet. Let's
temporarily revert these patches while we track down the root cause.
This commit squashes the reverts of:
76305b1 drm/i915: Calculate watermark configuration during atomic check (v2)
a4611e4 drm/i915: Don't set plane visible during HW readout if CRTC is off
a28170f drm/i915: Calculate ILK-style watermarks during atomic check (v3)
de4a9f8 drm/i915: Calculate pipe watermarks into CRTC state (v3)
de165e0 drm/i915: Refactor ilk_update_wm (v3)
Reference: http://lists.freedesktop.org/archives/intel-gfx/2015-October/077190.html
Cc: "Zanoni, Paulo R" <paulo.r.zanoni@intel.com>
Cc: "Vetter, Daniel" <daniel.vetter@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
During reset flow, ignore firmware errors detected prior
to the actual hardware reset as the recovery flow would
make additional unnecessary reset.
Signed-off-by: Hamad Kadmany <qca_hkadmany@qca.qualcomm.com>
Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
When firmware crashes, just before firmware recovery,
dump the firmware memory to a devcoredump device.
The resulting dump can be read from user space to be used
in offline crash analysis.
Signed-off-by: Lior David <liord@codeaurora.org>
Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
Situations observed when IP stack schedules lots of
frames for Tx while no connection (connection lost,
for example). In this case, dmesg bloated with error
message "FW not connected", printed for every frame.
Ratelimit this error message to avoid dmesg pollution.
Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
For the sniffer (monitor) mode, capture either control only or both
control and data PHY.
It used to be control only or data only PHY due to firmware
issues with configuration for PHY auto-detection; but now
it is resolved.
Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
BAR frames delivered to the host via Rx path; whole BAR frame
get delivered. Advance sequence in the reorder buffer and release
old frames, as per IEEE802.11 spec.
Firmware will reply to BAR, driver responsibility is only reorder
buffer management.
Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
Hostapd request disconnect for broadcast bssid when it
wants to disconnect all stations from the AP.
Detect this and really disconnect all connected stations.
Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
When allocating pmc descriptor, the structure is
initially created on stack and later copied to
the physical ring (device) memory. The descriptor
structure must be initialized to zero to avoid
garbage configuration, which may result in pmc
mechanism malfunctioning.
Signed-off-by: Vladimir Shulman <QCA_shulmanv@QCA.qualcomm.com>
Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
Print firmware and ucode assert codes when firmware crashed.
Signed-off-by: Lior David <liord@qti.qualcomm.com>
Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
On Rx, when invalid frame is received and dropped,
reaping of next frames from Rx ring is stopped.
This stops NAPI polling and re-enables the Rx interrupt.
However, in cases where no more frames received,
interrupt will not be triggered and rest of Rx frames
will not be processed.
Skip bad frames and continue to reap Rx packets when
such frames are encountered, and add statistics for
such frames for debug.
Signed-off-by: Hamad Kadmany <qca_hkadmany@qca.qualcomm.com>
Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
Fix compilation warning where CONFIG_PM defined while
CONFIG_PM_SLEEP is not defined
Report follows:
tree: git://github.com/kvalo/ath pending
head: 941145fc5e5afbb120271e5dfaf37213ddb55807
commit: df596be39294d9712e5d568063a48448031e0a9f [37/39] wil6210: system power management
config: xtensa-allyesconfig (attached as .config)
reproduce:
wget https://git.kernel.org/cgit/linux/kernel/git/wfg/lkp-tests.git/plain/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
git checkout df596be39294d9712e5d568063a48448031e0a9f
# save the attached .config to linux build tree
make.cross ARCH=xtensa
All warnings (new ones prefixed by >>):
>> drivers/net/wireless/ath/wil6210/pcie_bus.c:264:12: warning: 'wil6210_suspend' defined but not used [-Wunused-function]
static int wil6210_suspend(struct device *dev, bool is_runtime)
^
>> drivers/net/wireless/ath/wil6210/pcie_bus.c:291:12: warning: 'wil6210_resume' defined but not used [-Wunused-function]
static int wil6210_resume(struct device *dev, bool is_runtime)
^
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Vladimir Kondratiev <qca_vkondrat@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
The current number of spatial streams used by the client is advertised
as a separate IE in assoc request. Use this information to set
the NSS operating mode.
Fixes: 45c9abc059 ("ath10k: implement more versatile set_bitrate_mask").
Signed-off-by: Vivek Natarajan <nataraja@qti.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
After processing received packets from copy engine, host will allocate
new buffer and queue them back to copy engine ring for further
packet reception. On post rx processing path, skb allocation and
dma mapping are unnecessarily handled within ce_lock. This is affecting
peak throughput and also causing more CPU consumption. Optimize this
by acquiring ce_lock only when accessing copy engine ring and moving
skb allocation out of ce_lock.
In AP148 platform with QCA99x0 in conducted environment, UDP uplink peak
throughput is improved from ~1320 Mbps to ~1450 Mbps and TCP uplink peak
throughput is increased from ~1240 Mbps (70% host CPU load) to ~1300 Mbps
(71% CPU load). Similarly ~40Mbps improvement is observed in downlink
path.
Signed-off-by: Rajkumar Manoharan <rmanohar@qti.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
It is noticed that pci wakeup time is exceeding current timeout (10ms)
randomly which is tested on QCA988x. So, the wake up time is increased
to 30 ms and added debug prints to log total timeout.
Signed-off-by: Maharaja Kennadyrajan <c_mkenna@qti.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
When submitting semaphores in execlist mode the hang checker crashes in this
function because it is only runnable in ring submission mode. The reason this
is of particular interest to the TDR patch series is because we use semaphores
as a mean to induce hangs during testing (which is the recommended way to
induce hangs for gen8+). It's not clear how this is supposed to work in
execlist mode since:
1. This function requires a ring buffer.
2. Retrieving a ring buffer in execlist mode requires us to retrieve the
corresponding context, which we get from a request.
3. Retieving a request from the hang checker is not straight-forward since that
requires us to grab the struct_mutex in order to synchronize against the
request retirement thread.
4. Grabbing the struct_mutex from the hang checker is nothing that we will do
since that puts us at risk of deadlock since a hung thread might be holding the
struct_mutex already.
Therefore it's not obvious how we're supposed to deal with this. For now, we're
doing an early exit from this function, which avoids any kernel panic situation
when running our own internal TDR ULT.
* v2: (Chris Wilson)
Turned the execlist mode check into a ringbuffer NULL check to make it more
submission mode agnostic and less of a layering violation.
Signed-off-by: Tomas Elf <tomas.elf@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
commit e9f24d5fb7
Author: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Date: Mon Oct 5 13:26:36 2015 +0100
drm/i915: Clean up associated VMAs on context destruction
Introduced a wrong assumption that all contexts have a ppgtt
instance. This is not true when full PPGTT is not active so
remove the WARN_ON_ONCE from the context cleanup code.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Michel Thierry <michel.thierry@intel.com>
Reviewed-by: Michel Thierry <michel.thierry@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
There isn't an explicit stolen memory base register on gen2.
Some old comment in the i915 code suggests we should get it via
max_low_pfn_mapped, but that's clearly a bad idea on my MGM.
The e820 map in said machine looks like this:
[ 0.000000] BIOS-e820: [mem 0x0000000000000000-0x000000000009f7ff] usable
[ 0.000000] BIOS-e820: [mem 0x000000000009f800-0x000000000009ffff] reserved
[ 0.000000] BIOS-e820: [mem 0x00000000000ce000-0x00000000000cffff] reserved
[ 0.000000] BIOS-e820: [mem 0x00000000000dc000-0x00000000000fffff] reserved
[ 0.000000] BIOS-e820: [mem 0x0000000000100000-0x000000001f6effff] usable
[ 0.000000] BIOS-e820: [mem 0x000000001f6f0000-0x000000001f6f7fff] ACPI data
[ 0.000000] BIOS-e820: [mem 0x000000001f6f8000-0x000000001f6fffff] ACPI NVS
[ 0.000000] BIOS-e820: [mem 0x000000001f700000-0x000000001fffffff] reserved
[ 0.000000] BIOS-e820: [mem 0x00000000fec10000-0x00000000fec1ffff] reserved
[ 0.000000] BIOS-e820: [mem 0x00000000ffb00000-0x00000000ffbfffff] reserved
[ 0.000000] BIOS-e820: [mem 0x00000000fff00000-0x00000000ffffffff] reserved
That makes max_low_pfn_mapped = 1f6f0000, so assuming our stolen memory
would start there would place it on top of some ACPI memory regions.
So not a good idea as already stated.
The 9MB region after the ACPI regions at 0x1f700000 however looks
promising given that the macine reports the stolen memory size to be
8MB. Looking at the PGTBL_CTL register, the GTT entries are at offset
0x1fee00000, and given that the GTT entries occupy 128KB, it looks like
the stolen memory could start at 0x1f700000 and the GTT entries would
occupy the last 128KB of the stolen memory.
After some more digging through chipset documentation, I've determined
the BIOS first allocates space for something called TSEG (something to
do with SMM) from the top of memory, and then it allocates the graphics
stolen memory below that. Accordind to the chipset documentation TSEG
has a fixed size of 1MB on 855. So that explains the top 1MB in the
e820 region. And it also confirms that the GTT entries are in fact at
the end of the the stolen memory region.
Derive the stolen memory base address on gen2 the same as the BIOS does
(TOM-TSEG_SIZE-stolen_size). There are a few differences between the
registers on various gen2 chipsets, so a few different codepaths are
required.
865G is again bit more special since it seems to support enough memory
to hit 4GB address space issues. This means the PCI allocations will
also affect the location of the stolen memory. Fortunately there
appears to be the TOUD register which may give us the correct answer
directly. But the chipset docs are a bit unclear, so I'm not 100%
sure that the graphics stolen memory is always the last thing the
BIOS steals. Someone would need to verify it on a real system.
I tested this on the my 830 and 855 machines, and so far everything
looks peachy.
v2: Rewrite to use the TOM-TSEG_SIZE-stolen_size and TOUD methods
v3: Fix TSEG size for 830
v4: Add missing 'else' (Chris)
Tested-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Currently, when asynchronous transactions finish in error state and
retries, work scheduling and work running also continues. This
should be canceled at fatal error because it can cause endless loop.
This commit enables to cancel transferring MIDI messages when transactions
encounter fatal errors. This is achieved by setting error state.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Typically, the target devices have internal buffer to adjust output of
received MIDI messages for MIDI serial bus, while the capacity of the
buffer is limited. IEEE 1394 transactions can transfer more MIDI messages
than MIDI serial bus can. This can cause buffer over flow in device side.
This commit adds throttle to limit MIDI data rate by counting intervals
between two MIDI messages. Usual MIDI messages consists of two or three
bytes. This requires 1.302 to 1.953 mili-seconds interval between these
messages. This commit uses kernel monotonic time service to calculate the
time of next transaction.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Currently, when two MIDI trigger callbacks can be called immediately,
transactions for the second MIDI messages can be postpone till next trigger
callback. This is not good for real-time message transmission.
This commit schedules work again at response handling callback if the
MIDI substream still includes untransferred MIDI messages.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Currently, when waiting for a response, callers can start another
transaction by scheduling another work. This is not good for error
processing of transaction, especially the first response is too late.
This commit serialize request/response transactions, by adding one
boolean member to represent idling state.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Some models receive MIDI messages via IEEE 1394 asynchronous transactions.
In this case, MIDI messages are transferred in fixed-length payload. It's
nice that firewire-lib module has common helper functions.
This commit implements this idea. Each driver adds
'struct snd_fw_async_midi_port' in its instance structure. In probing,
it should call snd_fw_async_midi_port_init() to initialize the
structure with some parameters such as target address, the length
of payload in a transaction and a pointer for callback function
to fill the payload buffer. At 'struct snd_rawmidi_ops.trigger()'
callback, it should call 'snd_fw_async_midi_port_run()' to start
transactions. Each driver should ensure that the lifetime of MIDI
substream continues till calling 'snd_fw_async_midi_port_finish()'.
The helper functions support retries to transferring MIDI messages when
transmission errors occur. When transactions are successful, the helper
functions call 'snd_rawmidi_transmit_ack()' internally to consume MIDI
bytes in the buffer. Therefore, Each driver is expected to use
'snd_rawmidi_transmit_peek()' to tell the number of bytes to transfer to
return value of 'fill' callback.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
snd_seq_oss_readq_put_event() seems to be missing a memory barrier which
might cause the waker to not notice the waiter and miss sending a
wake_up as in the following figure.
snd_seq_oss_readq_put_event snd_seq_oss_readq_wait
------------------------------------------------------------------------
/* wait_event_interruptible_timeout */
/* __wait_event_interruptible_timeout */
/* ___wait_event */
for (;;) { prepare_to_wait_event(&wq, &__wait,
state);
spin_lock_irqsave(&q->lock, flags);
if (waitqueue_active(&q->midi_sleep))
/* The CPU might reorder the test for
the waitqueue up here, before
prior writes complete */
if ((q->qlen>0 || q->head==q->tail)
...
__ret = schedule_timeout(__ret)
if (q->qlen >= q->maxlen - 1) {
memcpy(&q->q[q->tail], ev, sizeof(*ev));
q->tail = (q->tail + 1) % q->maxlen;
q->qlen++;
------------------------------------------------------------------------
There are two other place in sound/core/seq/oss/ which have similar
code. The attached patch removes the call to waitqueue_active() leaving
just wake_up() behind. This fixes the problem because the call to
spin_lock_irqsave() in wake_up() will be an ACQUIRE operation.
I found this issue when I was looking through the linux source code
for places calling waitqueue_active() before wake_up*(), but without
preceding memory barriers, after sending a patch to fix a similar
issue in drivers/tty/n_tty.c (Details about the original issue can be
found here: https://lkml.org/lkml/2015/9/28/849).
Signed-off-by: Kosuke Tatsukawa <tatsu@ab.jp.nec.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
GCC is unable to properly optimize functions that have a very
short likely case and a longer and register-heavier cold part --
it fails to sink all of the register saving and stack frame
setup code into the unlikely part.
Help it out with syscall_return_slowpath() by splitting it into
two parts and inline the hot part.
Saves 6 cycles for compat syscalls.
Signed-off-by: Andy Lutomirski <luto@kernel.org>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Link: http://lkml.kernel.org/r/0f773a894ab15c589ac794c2d34ca6ba9b5335c9.1444091585.git.luto@kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
