Eric Laurent [Wed, 2 Mar 2016 17:54:57 +0000 (09:54 -0800)]
ALSA: compress: allow writes in SNDRV_PCM_STATE_PREPARED state
Allow writes in SNDRV_PCM_STATE_PREPARED state so that more
than one buffer fragment can be written from user space
before calling SNDRV_COMPRESS_START.
Thomas Klausner [Thu, 3 Mar 2016 15:28:22 +0000 (16:28 +0100)]
ALSA: uapi: Make alsa asound.h header more portable
This header is used as-is in the alsa-lib userland library,
which is portable to other operating systems.
For this reason, include linux/types.h only on Linux systems.
Add sys/ioctl.h for _IOR/_IOW/etc. (works at least on *BSD and Solaris).
Signed-off-by: Thomas Klausner <wiz@NetBSD.org> Signed-off-by: Takashi Iwai <tiwai@suse.de>
Libin Yang [Fri, 4 Mar 2016 06:33:43 +0000 (14:33 +0800)]
ALSA: hda - hdmi defer to register acomp eld notifier
Defer to register acomp eld notifier until hdmi audio driver
is fully ready.
After registering eld notifier, gfx driver can use this
callback function to notify audio driver the monitor
connection event. However this action may happen when
audio driver is adding the pins or doing other initialization.
This is not always safe, however. For example, using
per_pin->lock before the lock is initialized.
Let's register the eld notifier after the initialization is done.
Signed-off-by: Libin Yang <libin.yang@linux.intel.com> Signed-off-by: Takashi Iwai <tiwai@suse.de>
Libin Yang [Fri, 4 Mar 2016 06:33:06 +0000 (14:33 +0800)]
ALSA: hda - hdmi add wmb barrier for audio component
To make sure audio_ptr is set before intel_audio_codec_enable()
or intel_audio_codec_disable() calling pin_eld_notify(),
this patch adds wmb barrier to prevent optimizing.
Signed-off-by: Libin Yang <libin.yang@linux.intel.com> Signed-off-by: Takashi Iwai <tiwai@suse.de>
Simon South [Thu, 3 Mar 2016 04:10:44 +0000 (23:10 -0500)]
ALSA: hda - Fix mic issues on Acer Aspire E1-472
This patch applies the microphone-related fix created for the Acer
Aspire E1-572 to the E1-472 as well, as it uses the same Realtek ALC282
CODEC and demonstrates the same issues.
This patch allows an external, headset microphone to be used and limits
the gain on the (quite noisy) internal microphone.
Signed-off-by: Simon South <simon@simonsouth.com> Cc: <stable@vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai@suse.de>
Sudip Mukherjee [Wed, 2 Mar 2016 11:40:30 +0000 (17:10 +0530)]
ALSA: portman2x4: fix NULL pointer dereference
While registering pardev, the irq_func was also registered. As a result
when we tried to probe for the card, an interrupt was generated and in
the ISR we tried to dereference private_data. But private_data is still
NULL as we have not yet done portman_create(). Lets probe for the device
after card is created.
Fixes: e6a1b7e88046 ("ALSA: portman2x4 - use new parport device model") Signed-off-by: Sudip Mukherjee <sudip.mukherjee@codethink.co.uk> Signed-off-by: Takashi Iwai <tiwai@suse.de>
Takashi Iwai [Tue, 1 Mar 2016 17:30:18 +0000 (18:30 +0100)]
ALSA: seq: oss: Don't drain at closing a client
The OSS sequencer client tries to drain the pending events at
releasing. Unfortunately, as spotted by syzkaller fuzzer, this may
lead to an unkillable process state when the event has been queued at
the far future. Since the process being released can't be signaled
any longer, it remains and waits for the echo-back event in that far
future.
Back to history, the draining feature was implemented at the time we
misinterpreted POSIX definition for blocking file operation.
Actually, such a behavior is superfluous at release, and we should
just release the device as is instead of keeping it up forever.
This patch just removes the draining call that may block the release
for too long time unexpectedly.
Dennis Kadioglu [Tue, 1 Mar 2016 13:23:29 +0000 (14:23 +0100)]
ALSA: usb-audio: Add a quirk for Plantronics DA45
Plantronics DA45 does not support reading the sample rate which leads
to many lines of "cannot get freq at ep 0x4" and "cannot get freq at
ep 0x84". This patch adds the USB ID of the DA45 to quirks.c and
avoids those error messages.
Signed-off-by: Dennis Kadioglu <denk@post.com> Cc: <stable@vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai@suse.de>
Takashi Iwai [Mon, 29 Feb 2016 13:32:42 +0000 (14:32 +0100)]
ALSA: hdspm: Fix zero-division
HDSPM driver contains a code issuing zero-division potentially in
system sample rate ctl code. This patch fixes it by not processing
a zero or invalid rate value as a divisor, as well as excluding the
invalid value to be passed via the given ctl element.
Sudip Mukherjee [Mon, 29 Feb 2016 12:13:30 +0000 (17:43 +0530)]
ALSA: mts64: fix NULL pointer dereference
While registering pardev, the irq_func was also registered. As a
result when we tried to probe for the card, an interrupt was generated
and in the ISR we tried to dereference private_data. But private_data
is still NULL as we have not yet done snd_mts64_create(). Lets probe
for the card after mts64 is created.
Takashi Iwai [Sun, 28 Feb 2016 10:41:47 +0000 (11:41 +0100)]
ALSA: timer: Fix ioctls for X32 ABI
X32 ABI takes the 64bit timespec, thus the timer user status ioctl becomes
incompatible with IA32. This results in NOTTY error when the ioctl is
issued.
Meanwhile, this struct in X32 is essentially identical with the one in
X86-64, so we can just bypassing to the existing code for this
specific compat ioctl.
Takashi Iwai [Sun, 28 Feb 2016 10:36:14 +0000 (11:36 +0100)]
ALSA: timer: Fix broken compat timer user status ioctl
The timer user status compat ioctl returned the bogus struct used for
64bit architectures instead of the 32bit one. This patch addresses
it to return the proper struct.
Takashi Iwai [Sun, 28 Feb 2016 10:28:08 +0000 (11:28 +0100)]
ALSA: rawmidi: Fix ioctls X32 ABI
Like the previous fixes for ctl and PCM, we need a fix for
incompatible X32 ABI regarding the rawmidi: namely, struct
snd_rawmidi_status has the timespec, and the size and the alignment on
X32 differ from IA32.
Takashi Iwai [Sun, 28 Feb 2016 10:23:09 +0000 (11:23 +0100)]
ALSA: pcm: Fix ioctls for X32 ABI
X32 ABI uses the 64bit timespec in addition to 64bit alignment of
64bit values. This leads to incompatibilities in some PCM ioctls
involved with snd_pcm_channel_info, snd_pcm_status and
snd_pcm_sync_ptr structs. Fix the PCM compat ABI for these ioctls
like the previous commit for ctl API.
Takashi Iwai [Sat, 27 Feb 2016 16:52:42 +0000 (17:52 +0100)]
ALSA: ctl: Fix ioctls for X32 ABI
The X32 ABI takes the same alignment like x86-64, and this may result
in the incompatible struct size from ia32. Unfortunately, we hit this
in some control ABI: struct snd_ctl_elem_value differs between them
due to the position of 64bit variable array. This ends up with the
unknown ioctl (ENOTTY) error.
The fix is to add the compat entries for the new aligned struct.
Takashi Sakamoto [Sat, 27 Feb 2016 12:01:21 +0000 (21:01 +0900)]
ALSA: dice: drop duplex streams synchronization to transfer own time stamps
This commit drops implementation of duplex streams synchronization
from ALSA dice driver, due to a reason of hardware design. This patch
allows dice-based units to generate sounds correctly when isochronous
packet streaming starts at first time.
In IEC 61883-6:2005, CIP packetization layer for AM824 data format
utilizes the value of SYT field in CIP header of received packet for
a reference to phase lock loop. Figure 3 in clause 4.3 describes it.
The value is an offset from cycle_time field of every cycle start packet
from cycle master on IEEE 1394 bus. The time calculated with these two
fields is called as 'presentation timestamp' which represents the time
to play data included in the packet.
Although, this idea includes some problems due to accuracy of timekeep in
cycle master, accuracy of transmission of cycle start packet on the bus
with the other units, accuracy of sampling clock in data transmitter side
and accuracy of replay in data receiver side. In most case, these
accuracies somewhat worse because there's no such ideal hardwares in this
world.
For the issues, ASICs for Dice include Jitter Elimination Technologies
(JET) PLL. The PLL can handle several sources of clock and compensate it
with high-precision internal clock source. The sequence of value in syt
field of received AMDTP packets is one of the sources, therefore
transmitters on IEEE 1394 bus should transfer it.
On the other hand, current ALSA dice driver is programmed with a mode of
duplex streams with synchronization. In this mode, the driver outputs
packets after some incoming packets are handled, to re-use the value of
SYT field in incoming packets to the value for outgoing packets. This mode
is enabled when source signal of sampling clock is set to internal, and
this is a major use case. Thus, in most cases, the unit receives no packets
during a short time after packet streaming starts.
As long as I experienced, this causes the units to generate no sounds at
first time to receive packets. This issue occurs only with Dice II. I guess
this is due to a quirk of the PLL. In short, the PLL cannot generate firm
signals to ADCs/DACs or the other ICs when no packets are received in the
beginning of packet streaming. While, on second time or later, the unit
generates sound correctly. I guess that starting packet streaming at first
time sets the PLL correctly.
Well, still based on my hypothesis and no way to prove it, this commit
drops duplex streams synchronization from this driver. At least, the PLL
requires the sequence of value in SYT field of received AMDTP packets as
one of source of clock signals with internal clock source.
Ville Syrjälä [Fri, 26 Feb 2016 17:39:57 +0000 (19:39 +0200)]
ALSA: hda - Autosuspend controller after probe even if codecs are already suspended
azx_probe_continue() uses pm_runtime_put_noidle() to drop the rpm
usage_count, which means that if it's the last reference the
autosuspend of the controller won't actually happen. So if the codecs
autosuspend before the azx_probe_continue() drops the last
reference we'll fail to autosuspend the controller. This does happen
in practice, but not every time. As can be seen in [1] the controller
autosuspend attempt fails due to the usage_count when suspending the
codecs. A bit later we see the the contoller usage_count dropping to
zero without further attempts at autosuspend.
Fix the problem by using pm_runtime_put_autosuspend() instead, which
will kick off the autosuspend of the controller even if the codecs
are already asleep. As can be seen in [2] the controller autosuspend
still fails while suspending the codecs, but later on we see another
autosuspend attempt after dropping the usage_count to 0.
I was also a bit worried that there might still be a race between the
controller autosuspend and the rest of the code in azx_probe_continue().
So I also tried replacing the the put_noidle() with put_sync_suspend().
No explosions occurred, so I'm somewhat satisfied that there are no
serious problems in this area.
Takashi Iwai [Tue, 23 Feb 2016 14:54:47 +0000 (15:54 +0100)]
ALSA: hda - Loop interrupt handling until really cleared
Currently the interrupt handler of HD-audio driver assumes that no irq
update is needed while processing the irq. But in reality, it has
been confirmed that the HW irq is issued even during the irq
handling. Since we clear the irq status at the beginning, process the
interrupt, then exits from the handler, the lately issued interrupt is
left untouched without being properly processed.
This patch changes the interrupt handler code to loop over the
check-and-process. The handler tries repeatedly as long as the IRQ
status are turned on, and either stream or CORB/RIRB is handled.
For checking the stream handling, snd_hdac_bus_handle_stream_irq()
returns a value indicating the stream indices bits. Other than that,
the change is only in the irq handler itself.
Reported-by: Libin Yang <libin.yang@linux.intel.com> Cc: <stable@vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai@suse.de>
Takashi Iwai [Thu, 25 Feb 2016 13:31:59 +0000 (14:31 +0100)]
ALSA: hda - Fix headset support and noise on HP EliteBook 755 G2
HP EliteBook 755 G2 with ALC3228 (ALC280) codec [103c:221c] requires
the known fixup (ALC269_FIXUP_HEADSET_MIC) for making the headset mic
working. Also, it suffers from the loopback noise problem, so we
should disable aamix path as well.
ALSA: hda - Fixup speaker pass-through control for nid 0x14 on ALC225
On one of the machines we enable, we found that the actual speaker volume
did not always correspond to the volume set in alsamixer. This patch
fixes that problem.
This patch was orginally written by Kailang @ Realtek, I've rebased it
to fit sound git master.
Kai-Heng Feng [Thu, 25 Feb 2016 07:19:38 +0000 (15:19 +0800)]
ALSA: hda - Fixing background noise on Dell Inspiron 3162
After login to the desktop on Dell Inspiron 3162,
there's a very loud background noise comes from the builtin speaker.
The noise does not go away even if the speaker is muted.
The noise disappears after using the aamix fixup.
Codec: Realtek ALC3234
Address: 0
AFG Function Id: 0x1 (unsol 1)
Vendor Id: 0x10ec0255
Subsystem Id: 0x10280725
Revision Id: 0x100002
No Modem Function Group found
Takashi Sakamoto [Wed, 24 Feb 2016 00:26:33 +0000 (09:26 +0900)]
ALSA: oxfw: discontinue MIDI substream for scs1x at transaction failure
With a previous commit, ALSA oxfw driver retries transferring MIDI
messages at transaction failure for scs1x. On the other hand, there're
fatal transaction error. Then, no MIDI messages reach to the unit anymore.
In this case, MIDI substream should be terminated.
This commit stops MIDI transmission after the fatal error occurs.
Unfortunately, unlike ALSA PCM functionality, ALSA rawmidi core has no
feature to discontinue MIDI substream runtime in kernel side, thus this
commit just stops MIDI transmission without notifying it to userspace.
Takashi Sakamoto [Wed, 24 Feb 2016 00:26:32 +0000 (09:26 +0900)]
ALSA: oxfw: retry MIDI transferring for scs1x at transaction failure
Currently, ALSA oxfw driver has a TODO to retry MIDI transferring
at transaction failure.
This commit achieves it. Current implementation uses snd_rawmidi_transmit()
to transfer messages, thus the target MIDI messages are not in buffer when
transaction failure is detected. Although we cannot use a pair of
snd_rawmidi_transmit_peek() and snd_ramwidi_transmit_ack(), the
messages are still in scs1x specific structure and the data is available
for retries.
This commit adds a member to the structure for the length of buffered
messages, and uses the value again at retries.
Takashi Iwai [Wed, 17 Feb 2016 08:44:25 +0000 (09:44 +0100)]
ALSA: jack: Allow building the jack layer without input device
Since the recent integration of kctl jack and input jack layers, we
can basically build the jack layer even without input devices. That
is, the jack layer itself can be built with conditional to enable the
input device support or not, while the users may enable always
CONFIG_SND_JACK unconditionally.
For achieving it, this patch changes the following:
- A new Kconfig, CONFIG_SND_JACK_INPUT_DEV, was introduced to indicate
whether the jack layer supports the input device,
- A few items in snd_jack struct and relevant codes are conditionally
built upon CONFIG_SND_JACK_INPUT_DEV,
- The users of CONFIG_SND_JACK drop the messy dependency on
CONFIG_INPUT.
This change also automagically fixes a potential bug in HD-audio
driver Arnd reported, where the NULL or uninitialized jack instance is
dereferenced.
Takashi Iwai [Mon, 22 Feb 2016 14:18:13 +0000 (15:18 +0100)]
ALSA: hda - Apply clock gate workaround to Skylake, too
Some Skylake machines show the codec probe errors in certain
situations, e.g. HP Z240 desktop fails to probe the onboard Realtek
codec at reloading the snd-hda-intel module like:
snd_hda_intel 0000:00:1f.3: spurious response 0x200:0x2, last cmd=0x000000
snd_hda_intel 0000:00:1f.3: azx_get_response timeout, switching to polling mode: lastcmd=0x000f0000
snd_hda_intel 0000:00:1f.3: No response from codec, disabling MSI: last cmd=0x000f0000
snd_hda_intel 0000:00:1f.3: Codec #0 probe error; disabling it...
hdaudio hdaudioC0D2: no AFG or MFG node found
snd_hda_intel 0000:00:1f.3: no codecs initialized
Also, HP G470 G3 suffers from the similar problem, as reported in
bugzilla below. On this machine, the codec probe error appears even
at a fresh boot.
As Libin suggested, the same workaround used for Broxton in the commit
[6639484ddaf6: ALSA: hda - disable dynamic clock gating on Broxton
before reset] can be applied for Skylake in order to fix this problem.
The Intel HW team also confirmed that this is needed for SKL.
This patch makes the workaround applied to both SKL and BXT
platforms. The referred macros are moved and one superfluous macro
(IS_BROXTON()) is another one (IS_BXT()) as well.
Sudip Mukherjee [Mon, 22 Feb 2016 13:22:08 +0000 (18:52 +0530)]
ALSA: mts64: use new parport device model
Modify mts64 driver to use the new parallel port device model.
The advantage of using the device model is that the driver gets binded
to the hardware, we get the feature of hotplug, we can bind/unbind the
driver at runtime.
The changes are in the way the driver gets registered with the parallel
port subsystem and the temporary device to probe mts64 card is removed
and mts64_probe() is used in the probe callback.
Lukas Wunner [Thu, 14 Jan 2016 21:05:03 +0000 (22:05 +0100)]
ALSA: hda - Use acpi_dev_present()
Use shiny new acpi_dev_present() and remove all the boilerplate
to search for a particular ACPI device. No functional change.
Cf. 2d12b6b381ba ("ACPI / utils: Add acpi_dev_present()").
Takashi Sakamoto [Sat, 20 Feb 2016 07:41:01 +0000 (16:41 +0900)]
ALSA: fireworks: serialize transactions to update connections at bus reset
In IEC 61883-1, at bus-reset, applications can continue isochronous
streaming by updating connections. In ALSA fireworks driver, the
operation is executed in 'update' handler for bus driver.
The connection resources are also changed in process contexts of PCM/MIDI
applications. Therefore, bus-reset handling has race condition
against connection. Current ALSA fireworks driver has a bug for the
condition.
This commit fixes the bug, by expand critical section with mutex. As a
result, connection updating operation in bus-reset handler and connection
changing operation in process context are serialized.
Takashi Sakamoto [Sat, 20 Feb 2016 07:18:59 +0000 (16:18 +0900)]
ALSA: bebob: give up updating streams at bus reset handler
DM1000/DM1100/DM1500 chipsets transfer packets with discontinue value in
'dbc' field of CIP header. For ALSA bebob driver, this makes its bus-reset
handler meaningless, because the discontinuity is detected quite earlier
than executing the handler.
This commit gives up updating streams at the bus reset handler.
Takashi Sakamoto [Sat, 20 Feb 2016 07:18:57 +0000 (16:18 +0900)]
ALSA: bebob: move mutex from function callee to callers
Currently, critical section is protected by mutex in functions of
fireworks_stream.c. Callers increments/decrements substreams counter
before calling the functions. Moving mutex to the callers code allows
to change type of the substream counter from atomic_t to unsigned int.
This commit is a preparation for obsoleting usage of atomic_t for
substream counter.
Takashi Sakamoto [Sat, 20 Feb 2016 07:18:56 +0000 (16:18 +0900)]
ALSA: bebob: simplify bus-reset handling
At bus-reset, DM1000/DM1100/DM1500 chipsets transfer packets with
discontinuous value in 'dbc' field of CIP header. In this case, packet
streaming layer in firewire-lib module stops streaming and set XRUN to PCM
substream.
In ALSA, PCM applications are notified the XRUN status by the return value
of ALSA PCM interface. They can recover this state by executing
snd_pcm_prepare(), then PCM drivers' prepare handler is called, and start
new PCM substream. For ALSA BeBoB driver, the handler establishes new
connections and start new AMDTP streaming.
Unfortunately, neither the PCM applications nor the driver know the reason
of XRUN. The driver gets to know the reason when update handler is called
by IEEE 1394 bus driver. As long as I tested, the order of below events are
not fixed:
* Detecting packet discontinuity in tasklet context of OHCI 1394 driver
* Calling prepare handler in process context of ALSA PCM application
* Calling update handler in kthread context of IEEE 1394 bus driver
The unpredictable order is disadvantage for the driver to be compliant to
CMP. In IEC 61883-1, new CMP establish operations should be done 1 sec
(isoc_resource_delay) after bus-reset. Within 1 sec, CMP restore
operations are allowed. For this reason, in former commit ('b6bc812327aa:
ALSA: bebob/firewire-lib: Add a quirk for discontinuity at bus reset'),
the process context is forced to wait for executing update handler. The
process context wait for bus-reset up to 1 sec. This commit solves the
issue, while causes more disadvantages. For PCM applications, calling
snd_pcm_prepare() for recovering XRUN state takes more time and the driver
got a bit complicated code, while the recovery is not always successful.
As long as I tested, DM1000/DM1100/DM1500 and BeBoB firmware can allow
drivers to establish new connections just after bus reset. Furthermore,
any FCP transactions are handled correctly. Therefore, the driver don't
need to wait for bus reset handler for starting new streaming.
This commit removes the codes to reduce maintenance cost.
Libin Yang [Fri, 19 Feb 2016 07:42:06 +0000 (15:42 +0800)]
ALSA: hda - hdmi get jack from hda_jack_tbl when not dyn_pcm_assign
On Intel platform, if !dyn_pcm_assign, spec->pcm_rec[].jack is not
NULL even after snd_hda_jack_tbl_clear() is called to free snd_jack.
This may cause access invalid memory when calling snd_jack_report.
Fixes: 25e4abb33df3 ('ALSA: hda - hdmi jack created based on pcm') Signed-off-by: Libin Yang <libin.yang@linux.intel.com> Signed-off-by: Takashi Iwai <tiwai@suse.de>
Takashi Sakamoto [Wed, 17 Feb 2016 16:24:36 +0000 (01:24 +0900)]
ALSA: oxfw: use workqueue instead of tasklet for scs1x
This commit replaces tasklet with workqueue for scs1x functionality of
ALSA oxfw driver.
This driver transfers MIDI message specific for SCS.1m and SCS.1d. This
task is currently done in software IRQ context of tasklet. In a view of
system, this context is limited resources and some important drivers (at
least, more important than ALSA oxfw driver) use the context as its
bottom-harf.
If the work to transfer MIDI messages is done within a time, it's better
to use the other context for the work. Actually, with recent CPUs, the
work will be scheduled within a time. This is a reason of this commit.
Takashi Iwai [Wed, 17 Feb 2016 13:30:26 +0000 (14:30 +0100)]
ALSA: pcm: Fix rwsem deadlock for non-atomic PCM stream
A non-atomic PCM stream may take snd_pcm_link_rwsem rw semaphore twice
in the same code path, e.g. one in snd_pcm_action_nonatomic() and
another in snd_pcm_stream_lock(). Usually this is OK, but when a
write lock is issued between these two read locks, the problem
happens: the write lock is blocked due to the first reade lock, and
the second read lock is also blocked by the write lock. This
eventually deadlocks.
The reason is the way rwsem manages waiters; it's queued like FIFO, so
even if the writer itself doesn't take the lock yet, it blocks all the
waiters (including reads) queued after it.
As a workaround, in this patch, we replace the standard down_write()
with an spinning loop. This is far from optimal, but it's good
enough, as the spinning time is supposed to be relatively short for
normal PCM operations, and the code paths requiring the write lock
aren't called so often.
Sudip Mukherjee [Wed, 17 Feb 2016 12:56:27 +0000 (18:26 +0530)]
ALSA: portman2x4 - use new parport device model
Modify portman driver to use the new parallel port device model.
The advantage of using the device model is that the device gets binded
to the hardware, we get the feature of hotplug, we can bind/unbind
the driver at runtime.
The changes are in the way the driver gets registered with the
parallel port subsystem and the temporary device to probe portman card
is removed and portman_probe() is used in the probe callback.
Takashi Iwai [Tue, 16 Feb 2016 13:15:59 +0000 (14:15 +0100)]
ALSA: seq: Fix double port list deletion
The commit [7f0973e973cd: ALSA: seq: Fix lockdep warnings due to
double mutex locks] split the management of two linked lists (source
and destination) into two individual calls for avoiding the AB/BA
deadlock. However, this may leave the possible double deletion of one
of two lists when the counterpart is being deleted concurrently.
It ends up with a list corruption, as revealed by syzkaller fuzzer.
This patch fixes it by checking the list emptiness and skipping the
deletion and the following process.
Takashi Iwai [Mon, 15 Feb 2016 15:37:24 +0000 (16:37 +0100)]
ALSA: hda - Cancel probe work instead of flush at remove
The commit [991f86d7ae4e: ALSA: hda - Flush the pending probe work at
remove] introduced the sync of async probe work at remove for fixing
the race. However, this may lead to another hangup when the module
removal is performed quickly before starting the probe work, because
it issues flush_work() and it's blocked forever.
The workaround is to use cancel_work_sync() instead of flush_work()
there.
Fixes: 991f86d7ae4e ('ALSA: hda - Flush the pending probe work at remove') Cc: <stable@vger.kernel.org> # v3.17+ Signed-off-by: Takashi Iwai <tiwai@suse.de>
Takashi Iwai [Mon, 15 Feb 2016 15:20:24 +0000 (16:20 +0100)]
ALSA: seq: Fix leak of pool buffer at concurrent writes
When multiple concurrent writes happen on the ALSA sequencer device
right after the open, it may try to allocate vmalloc buffer for each
write and leak some of them. It's because the presence check and the
assignment of the buffer is done outside the spinlock for the pool.
The fix is to move the check and the assignment into the spinlock.
(The current implementation is suboptimal, as there can be multiple
unnecessary vmallocs because the allocation is done before the check
in the spinlock. But the pool size is already checked beforehand, so
this isn't a big problem; that is, the only possible path is the
multiple writes before any pool assignment, and practically seen, the
current coverage should be "good enough".)
Andrey Konovalov [Sat, 13 Feb 2016 08:08:06 +0000 (11:08 +0300)]
ALSA: usb-audio: avoid freeing umidi object twice
The 'umidi' object will be free'd on the error path by snd_usbmidi_free()
when tearing down the rawmidi interface. So we shouldn't try to free it
in snd_usbmidi_create() after having registered the rawmidi interface.
Takashi Iwai [Wed, 10 Feb 2016 11:47:03 +0000 (12:47 +0100)]
ALSA: timer: Call notifier in the same spinlock
snd_timer_notify1() is called outside the spinlock and it retakes the
lock after the unlock. This is rather racy, and it's safer to move
snd_timer_notify() call inside the main spinlock.
The patch also contains a slight refactoring / cleanup of the code.
Now all start/stop/continue/pause look more symmetric and a bit better
readable.
Takashi Sakamoto [Thu, 11 Feb 2016 11:18:38 +0000 (20:18 +0900)]
ALSA: dice: old firmware optimization for Dice notification
As long as I tested, Dice-based models produced by TC Electronic with
factory-configured settings transfer no notification within
ensure_phase_lock(). On the other hand, with upgraded firmwares, it
starts to transfer the notification. This seems to be a quirk of earlier
firmwares.
This commit ensures phase lock by reading a register after waiting for
the notification. Even if it's timed-out, ensure_phase_lock() return
success as long as the register has expected clock status.
Takashi Sakamoto [Thu, 11 Feb 2016 11:18:37 +0000 (20:18 +0900)]
ALSA: dice: change notification mask to detect lock status change
With former patchset, ALSA dice driver doesn't change clock parameters
anymore, while the driver still touch clock configuration for phase lock.
Although the locking status is in Dice notification, the driver doesn't
detect it. Usually, this causes no issues because in most case
NOTIFY_LOCK_CHG notification transfers after NOTIFY_CLOCK_ACCEPTED
notification, while it's better to detect locking status.
This commit changes notification mask just to detect lock status change.
Takashi Iwai [Fri, 12 Feb 2016 08:48:51 +0000 (09:48 +0100)]
Merge tag 'asoc-fix-v4.5-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound into for-linus
ASoC: Fixes for v4.5
A rather large batch of fixes here, almost all in the Intel driver.
The changes that got merged in this merge window for Skylake were rather
large and as well as issues that you'd expect in a large block of new
code there were some problems created for older processors which needed
fixing up. Things are largely settling down now hopefully.
ASoC: arizona: fref must be limited in pseudo-fractional mode
When the FLL is in pseudo-fractional mode there is an additional
limit on fref based on the fratio, to prevent aliasing around the
Nyquist frequency. If fref exceeds this limit the refclk divider
must be increased and the calculation tried again until a suitable
combination of fref and fratio is found or we have to fall back to
integer mode.
This patch also adds some debug log prints around this code.
Signed-off-by: Richard Fitzgerald <rf@opensource.wolfsonmicro.com> Signed-off-by: Mark Brown <broonie@kernel.org>
Pascal Huerst [Wed, 10 Feb 2016 14:59:28 +0000 (15:59 +0100)]
ASoC: sigmadsp: Fix missleading return value
Forwarding the return value of i2c_master_send, leads to errors
later on, since i2c_master_send returns the number of bytes
transmittet. Check for ret < 0 instead and return 0 otherwise.
Signed-off-by: Pascal Huerst <pascal.huerst@gmail.com> Acked-by: Lars-Peter Clausen <lars@metafoo.de> Signed-off-by: Mark Brown <broonie@kernel.org>
Takashi Iwai [Wed, 10 Feb 2016 10:53:30 +0000 (11:53 +0100)]
ALSA: timer: Protect the whole snd_timer_close() with open race
In order to make the open/close more robust, widen the register_mutex
protection over the whole snd_timer_close() function. Also, the close
procedure is slightly shuffled to be in the safer order, as well as a
few code refactoring.
Takashi Iwai [Mon, 8 Feb 2016 16:26:58 +0000 (17:26 +0100)]
ALSA: timer: Fix race at concurrent reads
snd_timer_user_read() has a potential race among parallel reads, as
qhead and qused are updated outside the critical section due to
copy_to_user() calls. Move them into the critical section, and also
sanitize the relevant code a bit.
ALSA: dice: ensure phase lock before starting streaming
In former commits, probing process has no need to set sampling transfer
frequency. Although it's OK to drop a function to change the frequency
from this module, some models require it before streaming. This seems to
be due to phase lock of clock source.
This commit moves the function from transaction layer to stream layer, and
rename it according to the purpose.
Dice interface design doesn't allow drivers to read supported combination
between sampling transfer frequencies and the number of Multi bit linear
audio data channels. Due to the design, ALSA dice driver changes current
sampling transfer frequency to generate cache of the combinations at
device probing processing.
Although, this idea is worse because ALSA dice driver changes the state of
clock. This is not what users want when they save favorite configuration
to the device in advance.
Furthermore, there's a possibility that the format of data block is decided
not only according to current sampling transfer frequency, but also the
other factors, i.e. data format for digital interface. It's not good to
generate channel cache according to the sampling transfer frequency only.
This commit purges processing cache data and related structure members. As
a result, users must set preferable sampling transfer frequency before
using ALSA PCM applications, as long as they want to start any PCM
substreams at the rate except for current one.
ALSA: dice: get the number of MBLA data channel at opening PCM substream
This commit is a preparation to remove members related to channel cache
for the number of channels for multi bit linear audio data and MIDI
ports. This commit changes the way to get the number of multi bit linear
audio data channel. It's directly retrieved by asynchronous transactions
to some registers.
ALSA: dice: add MIDI ports according to current number of MIDI substreams
This commit changes the way to add ALSA MIDI ports. This driver read the
number of multiplexed MIDI substreams from hardware register, then adds the
same number of ALSA MIDI ports. This commit is based on my assumption that
the number is fixed at all of supported sampling transfer frequency.
ALSA: dice: limit stream to current sampling transfer frequency.
In previous commit, ALSA Dice driver limits PCM substreams at current
sampling transfer frequency and current number of Multi bit linear audio
data channel. Thus, the driver has no need to start AMDTP streams at
the other sampling transfer frequency except for current one. This is due
to Dice interface design.
This commit limits AMDTP stream at current sampling transfer frequency,
according to the design.
ALSA: dice: limit to current sampling transfer frequency
ALSA PCM core has a functionality for rule of PCM substream parameters.
Typically, when userspace opens PCM character devices, each driver adds
its own rules to PCM substream according to design of hardware. When the
userspace executes hw_params ioctl with favorite parameters, the actual
parameters are calculated according to the rules and the given parameters.
Then, the result is returned to userspace.
Currently, ALSA Dice driver has the rule between channels and rates, while
Dice interface design doesn't allow drivers to retrieve all of the
combinations. Dice drivers are just allowed to get current sampling
transfer frequency and the number of multi bit linear audio data channels
in an data block of an AMDTP packet.
This commit purges the rule, and limit PCM substreams to current sampling
transfer frequency, following to the interface design.
Takashi Iwai [Tue, 9 Feb 2016 09:23:52 +0000 (10:23 +0100)]
ALSA: hda - Fix bad dereference of jack object
The hda_jack_tbl entries are managed by snd_array for allowing
multiple jacks. It's good per se, but the problem is that struct
hda_jack_callback keeps the hda_jack_tbl pointer. Since snd_array
doesn't preserve each pointer at resizing the array, we can't keep the
original pointer but have to deduce the pointer at each time via
snd_array_entry() instead. Actually, this resulted in the deference
to the wrong pointer on codecs that have many pins such as CS4208.
This patch replaces the pointer to the NID value as the search key.
As an unexpected good side effect, this even simplifies the code, as
only NID is needed in most cases.
Takashi Iwai [Tue, 9 Feb 2016 11:02:32 +0000 (12:02 +0100)]
ALSA: timer: Fix race between stop and interrupt
A slave timer element also unlinks at snd_timer_stop() but it takes
only slave_active_lock. When a slave is assigned to a master,
however, this may become a race against the master's interrupt
handling, eventually resulting in a list corruption. The actual bug
could be seen with a syzkaller fuzzer test case in BugLink below.
As a fix, we need to take timeri->timer->lock when timer isn't NULL,
i.e. assigned to a master, while the assignment to a master itself is
protected by slave_active_lock.
Takashi Iwai [Mon, 8 Feb 2016 16:36:25 +0000 (17:36 +0100)]
ALSA: timer: Fix wrong instance passed to slave callbacks
In snd_timer_notify1(), the wrong timer instance was passed for slave
ccallback function. This leads to the access to the wrong data when
an incompatible master is handled (e.g. the master is the sequencer
timer and the slave is a user timer), as spotted by syzkaller fuzzer.
DPCM driver is recommended for BYT, CHT based platforms, so if
CONFIG_SND_SST_IPC_ACPI is selected then don't compile the BYT
Device IDs in common ACPI driver to avoid probe conflicts.
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Acked-by: Jie Yang <yang.jie@intel.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com> Signed-off-by: Mark Brown <broonie@kernel.org>
Vinod Koul [Mon, 8 Feb 2016 05:15:37 +0000 (10:45 +0530)]
ASoC: Intel: Create independent acpi match module
The ACPI match module is common to all three drivers, HSW, SKL
and Atom-DPCM driver. But Atom-DPCM driver does not use common
sst code so we cannot include the common SST module in Atom-DPCM
driver.
So the solution is to have a independent sst-match-acpi module
which helps in matching for all the three drivers. Now all driver
can be inbuilt in a single image
This patch really fixes the regression introduced by the
commit 95f098014815 ("ASoC: Intel: Move apci find machine routines")
Acked-by: Jie Yang <yang.jie@intel.com> Acked-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com> Signed-off-by: Mark Brown <broonie@kernel.org>
Vinod Koul [Mon, 8 Feb 2016 05:15:36 +0000 (10:45 +0530)]
ASoC: Intel: Revert "ASoC: Intel: fix ACPI probe regression with Atom DPCM driver"
This reverts commit dc901a354171 ("ASoC: Intel: fix ACPI probe
regression with Atom DPCM driver") as the fix prevented the probe
on HSW/BDW if Atom-DPCM was selected
Acked-by: Jie Yang <yang.jie@intel.com> Acked-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Signed-off-by: Vinod Koul <vinod.koul@intel.com> Signed-off-by: Mark Brown <broonie@kernel.org>
The AU1X00 AC97 ALSA driver was deprecated in commit 7137c6bcb7ff ("ALSA:
deprecate MIPS AU1X00 AC97 driver") in favor of the newer and better ASoC
driver for the same hardware. This was almost 5 years ago and this driver
has not been in use in the mainline kernel since, it should be safe to
remove it at this point.
Takashi Iwai [Tue, 2 Feb 2016 14:27:36 +0000 (15:27 +0100)]
ALSA: dummy: Implement timer backend switching more safely
Currently the selected timer backend is referred at any moment from
the running PCM callbacks. When the backend is switched, it's
possible to lead to inconsistency from the running backend. This was
pointed by syzkaller fuzzer, and the commit [7ee96216c31a: ALSA:
dummy: Disable switching timer backend via sysfs] disabled the dynamic
switching for avoiding the crash.
This patch improves the handling of timer backend switching. It keeps
the reference to the selected backend during the whole operation of an
opened stream so that it won't be changed by other streams.
Together with this change, the hrtimer parameter is reenabled as
writable now.
NOTE: this patch also turned out to fix the still remaining race.
Namely, ops was still replaced dynamically at dummy_pcm_open:
static int dummy_pcm_open(struct snd_pcm_substream *substream)
{
....
dummy->timer_ops = &dummy_systimer_ops;
if (hrtimer)
dummy->timer_ops = &dummy_hrtimer_ops;
Since dummy->timer_ops is common among all streams, and when the
replacement happens during accesses of other streams, it may lead to a
crash. This was actually triggered by syzkaller fuzzer and KASAN.
This patch rewrites the code not to use the ops shared by all streams
any longer, too.
Linus Torvalds [Sun, 7 Feb 2016 23:23:20 +0000 (15:23 -0800)]
Merge tag 'armsoc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc
Pull ARM SoC fixes from Olof Johansson:
"The first real batch of fixes for this release cycle, so there are a
few more than usual.
Most of these are fixes and tweaks to board support (DT bugfixes,
etc). I've also picked up a couple of small cleanups that seemed
innocent enough that there was little reason to wait (const/
__initconst and Kconfig deps).
Quite a bit of the changes on OMAP were due to fixes to no longer
write to rodata from assembly when ARM_KERNMEM_PERMS was enabled, but
there were also other fixes.
Kirkwood had a bunch of gpio fixes for some boards. OMAP had RTC
fixes on OMAP5, and Nomadik had changes to MMC parameters in DT.
All in all, mostly the usual mix of various fixes"
* tag 'armsoc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc: (46 commits)
ARM: multi_v7_defconfig: enable DW_WATCHDOG
ARM: nomadik: fix up SD/MMC DT settings
ARM64: tegra: Add chosen node for tegra132 norrin
ARM: realview: use "depends on" instead of "if" after prompt
ARM: tango: use "depends on" instead of "if" after prompt
ARM: tango: use const and __initconst for smp_operations
ARM: realview: use const and __initconst for smp_operations
bus: uniphier-system-bus: revive tristate prompt
arm64: dts: Add missing DMA Abort interrupt to Juno
bus: vexpress-config: Add missing of_node_put
ARM: dts: am57xx: sbc-am57x: correct Eth PHY settings
ARM: dts: am57xx: cl-som-am57x: fix CPSW EMAC pinmux
ARM: dts: am57xx: sbc-am57x: fix UART3 pinmux
ARM: dts: am57xx: cl-som-am57x: update SPI Flash frequency
ARM: dts: am57xx: cl-som-am57x: set HOST mode for USB2
ARM: dts: am57xx: sbc-am57x: fix SB-SOM EEPROM I2C address
ARM: dts: LogicPD Torpedo: Revert Duplicative Entries
ARM: dts: am437x: pixcir_tangoc: use correct flags for irq types
ARM: dts: am4372: fix irq type for arm twd and global timer
ARM: dts: at91: sama5d4 xplained: fix phy0 IRQ type
...
Linus Torvalds [Sun, 7 Feb 2016 06:14:46 +0000 (22:14 -0800)]
Merge tag 'usb-4.5-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb
Pull USB fixes from Greg KH:
"Here are some USB fixes for 4.5-rc3.
The usual, xhci fixes for reported issues, combined with some small
gadget driver fixes, and a MAINTAINERS file update. All have been in
linux-next with no reported issues"
* tag 'usb-4.5-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb:
xhci: harden xhci_find_next_ext_cap against device removal
xhci: Fix list corruption in urb dequeue at host removal
usb: host: xhci-plat: fix NULL pointer in probe for device tree case
usb: xhci-mtk: fix AHB bus hang up caused by roothubs polling
usb: xhci-mtk: fix bpkts value of LS/HS periodic eps not behind TT
usb: xhci: apply XHCI_PME_STUCK_QUIRK to Intel Broxton-M platforms
usb: xhci: set SSIC port unused only if xhci_suspend succeeds
usb: xhci: add a quirk bit for ssic port unused
usb: xhci: handle both SSIC ports in PME stuck quirk
usb: dwc3: gadget: set the OTG flag in dwc3 gadget driver.
Revert "xhci: don't finish a TD if we get a short-transfer event mid TD"
MAINTAINERS: fix my email address
usb: dwc2: Fix probe problem on bcm2835
Revert "usb: dwc2: Move reset into dwc2_get_hwparams()"
usb: musb: ux500: Fix NULL pointer dereference at system PM
usb: phy: mxs: declare variable with initialized value
usb: phy: msm: fix error handling in probe.
Linus Torvalds [Sun, 7 Feb 2016 06:13:16 +0000 (22:13 -0800)]
Merge tag 'staging-4.5-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging
Pull staging and IIO driver fixes from Greg KH:
"Here are some IIO and staging driver fixes for 4.5-rc3.
All of them, except one, are for IIO drivers, and one is for a speakup
driver fix caused by some earlier patches, to resolve a reported build
failure"
* tag 'staging-4.5-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging:
Staging: speakup: Fix allyesconfig build on mn10300
iio: dht11: Use boottime
iio: ade7753: avoid uninitialized data
iio: pressure: mpl115: fix temperature offset sign
iio: imu: Fix dependencies for !HAS_IOMEM archs
staging: iio: Fix dependencies for !HAS_IOMEM archs
iio: adc: Fix dependencies for !HAS_IOMEM archs
iio: inkern: fix a NULL dereference on error
iio:adc:ti_am335x_adc Fix buffered mode by identifying as software buffer.
iio: light: acpi-als: Report data as processed
iio: dac: mcp4725: set iio name property in sysfs
iio: add HAS_IOMEM dependency to VF610_ADC
iio: add IIO_TRIGGER dependency to STK8BA50
iio: proximity: lidar: correct return value
iio-light: Use a signed return type for ltr501_match_samp_freq()
Linus Torvalds [Sat, 6 Feb 2016 04:20:07 +0000 (20:20 -0800)]
Merge branch 'akpm' (patches from Andrew)
Merge fixes from Andrew Morton:
"22 fixes"
* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (22 commits)
epoll: restrict EPOLLEXCLUSIVE to POLLIN and POLLOUT
radix-tree: fix oops after radix_tree_iter_retry
MAINTAINERS: trim the file triggers for ABI/API
dax: dirty inode only if required
thp: make deferred_split_scan() work again
mm: replace vma_lock_anon_vma with anon_vma_lock_read/write
ocfs2/dlm: clear refmap bit of recovery lock while doing local recovery cleanup
um: asm/page.h: remove the pte_high member from struct pte_t
mm, hugetlb: don't require CMA for runtime gigantic pages
mm/hugetlb: fix gigantic page initialization/allocation
mm: downgrade VM_BUG in isolate_lru_page() to warning
mempolicy: do not try to queue pages from !vma_migratable()
mm, vmstat: fix wrong WQ sleep when memory reclaim doesn't make any progress
vmstat: make vmstat_update deferrable
mm, vmstat: make quiet_vmstat lighter
mm/Kconfig: correct description of DEFERRED_STRUCT_PAGE_INIT
memblock: don't mark memblock_phys_mem_size() as __init
dump_stack: avoid potential deadlocks
mm: validate_mm browse_rb SMP race condition
m32r: fix build failure due to SMP and MMU
...
Linus Torvalds [Sat, 6 Feb 2016 03:52:57 +0000 (19:52 -0800)]
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client
Pull Ceph fixes from Sage Weil:
"We have a few wire protocol compatibility fixes, ports of a few recent
CRUSH mapping changes, and a couple error path fixes"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client:
libceph: MOSDOpReply v7 encoding
libceph: advertise support for TUNABLES5
crush: decode and initialize chooseleaf_stable
crush: add chooseleaf_stable tunable
crush: ensure take bucket value is valid
crush: ensure bucket id is valid before indexing buckets array
ceph: fix snap context leak in error path
ceph: checking for IS_ERR instead of NULL
Linus Torvalds [Sat, 6 Feb 2016 03:38:15 +0000 (19:38 -0800)]
Merge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux
Pull drm fixes from Dave Airlie:
"Fixes all over the place:
- amdkfd: two static checker fixes
- mst: a bunch of static checker and spec/hw interaction fixes
- amdgpu: fix Iceland hw properly, and some fiji bugs, along with
some write-combining fixes.
- exynos: some regression fixes
- adv7511: fix some EDID reading issues"
* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux: (38 commits)
drm/dp/mst: deallocate payload on port destruction
drm/dp/mst: Reverse order of MST enable and clearing VC payload table.
drm/dp/mst: move GUID storage from mgr, port to only mst branch
drm/dp/mst: change MST detection scheme
drm/dp/mst: Calculate MST PBN with 31.32 fixed point
drm: Add drm_fixp_from_fraction and drm_fixp2int_ceil
drm/mst: Add range check for max_payloads during init
drm/mst: Don't ignore the MST PBN self-test result
drm: fix missing reference counting decrease
drm/amdgpu: disable uvd and vce clockgating on Fiji
drm/amdgpu: remove exp hardware support from iceland
drm/amdgpu: load MEC ucode manually on iceland
drm/amdgpu: don't load MEC2 on topaz
drm/amdgpu: drop topaz support from gmc8 module
drm/amdgpu: pull topaz gmc bits into gmc_v7
drm/amdgpu: The VI specific EXE bit should only apply to GMC v8.0 above
drm/amdgpu: iceland use CI based MC IP
drm/amdgpu: move gmc7 support out of CIK dependency
drm/amdgpu/gfx7: enable cp inst/reg error interrupts
drm/amdgpu/gfx8: enable cp inst/reg error interrupts
...
Linus Torvalds [Sat, 6 Feb 2016 02:11:23 +0000 (18:11 -0800)]
Merge tag 'pm+acpi-4.5-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Pull power management and ACPI fixes from Rafael Wysocki:
"These are: a fix for a recently introduced false-positive warnings
about PM domain pointers being changed inappropriately (harmless but
annoying), an MCH size workaround quirk for one more platform, a
compiler warning fix (generic power domains framework), an ACPI LPSS
(Intel SoCs) driver fixup and a cleanup of the ACPI CPPC core code.
Specifics:
- PM core fix to avoid false-positive warnings generated when the
pm_domain field is cleared for a device that appears to be bound to
a driver (Rafael Wysocki).
- New MCH size workaround quirk for Intel Haswell-ULT (Josh Boyer).
- Fix for an "unused function" compiler warning in the generic power
domains framework (Ulf Hansson).
- Fixup for the ACPI driver for Intel SoCs (acpi-lpss) to set the PM
domain pointer of a device properly in one place that was
overlooked by a recent PM core update (Andy Shevchenko).
- Removal of a redundant function declaration in the ACPI CPPC core
code (Timur Tabi)"
* tag 'pm+acpi-4.5-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
PM: Avoid false-positive warnings in dev_pm_domain_set()
PM / Domains: Silence compiler warning for an unused function
ACPI / CPPC: remove redundant mbox_send_message() declaration
ACPI / LPSS: set PM domain via helper setter
PNP: Add Haswell-ULT to Intel MCH size workaround
Jason Baron [Fri, 5 Feb 2016 23:37:04 +0000 (15:37 -0800)]
epoll: restrict EPOLLEXCLUSIVE to POLLIN and POLLOUT
In the current implementation of the EPOLLEXCLUSIVE flag (added for
4.5-rc1), if epoll waiters create different POLL* sets and register them
as exclusive against the same target fd, the current implementation will
stop waking any further waiters once it finds the first idle waiter.
This means that waiters could miss wakeups in certain cases.
For example, when we wake up a pipe for reading we do:
wake_up_interruptible_sync_poll(&pipe->wait, POLLIN | POLLRDNORM); So if
one epoll set or epfd is added to pipe p with POLLIN and a second set
epfd2 is added to pipe p with POLLRDNORM, only epfd may receive the
wakeup since the current implementation will stop after it finds any
intersection of events with a waiter that is blocked in epoll_wait().
We could potentially address this by requiring all epoll waiters that
are added to p be required to pass the same set of POLL* events. IE the
first EPOLL_CTL_ADD that passes EPOLLEXCLUSIVE establishes the set POLL*
flags to be used by any other epfds that are added as EPOLLEXCLUSIVE.
However, I think it might be somewhat confusing interface as we would
have to reference count the number of users for that set, and so
userspace would have to keep track of that count, or we would need a
more involved interface. It also adds some shared state that we'd have
store somewhere. I don't think anybody will want to bloat
__wait_queue_head for this.
I think what we could do instead, is to simply restrict EPOLLEXCLUSIVE
such that it can only be specified with EPOLLIN and/or EPOLLOUT. So
that way if the wakeup includes 'POLLIN' and not 'POLLOUT', we can stop
once we hit the first idle waiter that specifies the EPOLLIN bit, since
any remaining waiters that only have 'POLLOUT' set wouldn't need to be
woken. Likewise, we can do the same thing if 'POLLOUT' is in the wakeup
bit set and not 'POLLIN'. If both 'POLLOUT' and 'POLLIN' are set in the
wake bit set (there is at least one example of this I saw in fs/pipe.c),
then we just wake the entire exclusive list. Having both 'POLLOUT' and
'POLLIN' both set should not be on any performance critical path, so I
think that's ok (in fs/pipe.c its in pipe_release()). We also continue
to include EPOLLERR and EPOLLHUP by default in any exclusive set. Thus,
the user can specify EPOLLERR and/or EPOLLHUP but is not required to do
so.
Since epoll waiters may be interested in other events as well besides
EPOLLIN, EPOLLOUT, EPOLLERR and EPOLLHUP, these can still be added by
doing a 'dup' call on the target fd and adding that as one normally
would with EPOLL_CTL_ADD. Since I think that the POLLIN and POLLOUT
events are what we are interest in balancing, I think that the 'dup'
thing could perhaps be added to only one of the waiter threads.
However, I think that EPOLLIN, EPOLLOUT, EPOLLERR and EPOLLHUP should be
sufficient for the majority of use-cases.
Since EPOLLEXCLUSIVE is intended to be used with a target fd shared
among multiple epfds, where between 1 and n of the epfds may receive an
event, it does not satisfy the semantics of EPOLLONESHOT where only 1
epfd would get an event. Thus, it is not allowed to be specified in
conjunction with EPOLLEXCLUSIVE.
EPOLL_CTL_MOD is also not allowed if the fd was previously added as
EPOLLEXCLUSIVE. It seems with the limited number of flags to not be as
interesting, but this could be relaxed at some further point.
Signed-off-by: Jason Baron <jbaron@akamai.com> Tested-by: Madars Vitolins <m@silodev.com> Cc: Michael Kerrisk <mtk.manpages@gmail.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Al Viro <viro@ftp.linux.org.uk> Cc: Eric Wong <normalperson@yhbt.net> Cc: Jonathan Corbet <corbet@lwn.net> Cc: Andy Lutomirski <luto@amacapital.net> Cc: Hagen Paul Pfeifer <hagen@jauu.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Helper radix_tree_iter_retry() resets next_index to the current index.
In following radix_tree_next_slot current chunk size becomes zero. This
isn't checked and it tries to dereference null pointer in slot.
Tagged iterator is fine because retry happens only at slot 0 where tag
bitmask in iter->tags is filled with single bit.
Fixes: 46437f9a554f ("radix-tree: fix race in gang lookup") Signed-off-by: Konstantin Khlebnikov <koct9i@gmail.com> Cc: Matthew Wilcox <willy@linux.intel.com> Cc: Hugh Dickins <hughd@google.com> Cc: Ohad Ben-Cohen <ohad@wizery.com> Cc: Jeremiah Mahler <jmmahler@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Commit ea8f8fc8631 ("MAINTAINERS: add linux-api for review of API/ABI
changes") added file triggers for various paths that likely indicated
API/ABI changes. However, catching all changes in Documentation/ABI/
and include/uapi/ produces a large volume of mail to linux-api, rather
than only API/ABI changes. Drop those two entries, but leave
include/linux/syscalls.h and kernel/sys_ni.c to catch syscall-related
changes.
Dmitry Monakhov [Fri, 5 Feb 2016 23:36:55 +0000 (15:36 -0800)]
dax: dirty inode only if required
Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org> Reviewed-by: Jan Kara <jack@suse.cz> Reviewed-by: Ross Zwisler <ross.zwisler@linux.intel.com> Cc: Matthew Wilcox <matthew.r.wilcox@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
We need to iterate over split_queue, not local empty list to get
anything split from the shrinker.
Fixes: e3ae19535c66 ("thp: limit number of object to scan on deferred_split_scan()") Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
projects / linux-beck.git / log