sfrench/cifs-2.6.git
3 years agoMerge branch 'for-linus' into for-next
Takashi Iwai [Mon, 14 Dec 2020 08:05:24 +0000 (09:05 +0100)]
Merge branch 'for-linus' into for-next

Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda/ca0132 - Add ZxR surround DAC setup.
Connor McAdams [Fri, 11 Dec 2020 22:55:03 +0000 (17:55 -0500)]
ALSA: hda/ca0132 - Add ZxR surround DAC setup.

Add pre-dsp download initialization for the DAC's used in the surround
sound configuration. Fixes issues of no audio on surround channels.

Fixes: 2e492b8ee5da8 ("ALSA: hda/ca0132 - Add ZxR init commands")
Signed-off-by: Connor McAdams <conmanx360@gmail.com>
Link: https://lore.kernel.org/r/20201211225504.4508-2-conmanx360@gmail.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda/ca0132 - Add 8051 PLL write helper functions.
Connor McAdams [Fri, 11 Dec 2020 22:55:02 +0000 (17:55 -0500)]
ALSA: hda/ca0132 - Add 8051 PLL write helper functions.

Add helper functions for the 8051 PLL PMU write verbs.

Signed-off-by: Connor McAdams <conmanx360@gmail.com>
Link: https://lore.kernel.org/r/20201211225504.4508-1-conmanx360@gmail.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda/hdmi: packet buffer index must be set before reading value
Kai Vehmanen [Fri, 11 Dec 2020 13:16:13 +0000 (15:16 +0200)]
ALSA: hda/hdmi: packet buffer index must be set before reading value

The check for infoframe transmit status in hdmi_infoframe_uptodate()
makes the assumption that packet buffer index is set to zero.

Align code with specification and explicitly set the index before
AC_VERB_GET_HDMI_DIP_XMIT. The packet index setting affects both
DIP-Data and DIP-XmitCtrl verbs.

There are no known cases where the old implementation has caused driver
to work incorrectly. This change is purely based on code review against
the specification (HDA spec rev1.0a).

Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Link: https://lore.kernel.org/r/20201211131613.3271407-1-kai.vehmanen@linux.intel.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Fix control 'access overflow' errors from chmap
Takashi Iwai [Fri, 11 Dec 2020 13:00:48 +0000 (14:00 +0100)]
ALSA: usb-audio: Fix control 'access overflow' errors from chmap

The current channel-map control implementation in USB-audio driver may
lead to an error message like
  "control 3:0:0:Playback Channel Map:0: access overflow"
when CONFIG_SND_CTL_VALIDATION is set.  It's because the chmap get
callback clears the whole array no matter which count is set, and
rather the false-positive detection.

This patch fixes the problem by clearing only the needed array range
at usb_chmap_ctl_get().

Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/20201211130048.6358-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda/hdmi: always print pin NIDs as hexadecimal
Kai Vehmanen [Fri, 11 Dec 2020 12:45:47 +0000 (14:45 +0200)]
ALSA: hda/hdmi: always print pin NIDs as hexadecimal

The debug prints from patch_hdmi.c are not aligned with HDA common code
in hda_codec.c nor with other HDA codec drivers.

To align with rest of the codebase, use hexadecimal formatting whenever
printing value of a HDA NID. Also refer to NIDs with capital letters in
traces as is done other modules. This presentation is also aligned with
the formatting used in HDA codec procfs entry.

Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Link: https://lore.kernel.org/r/20201211124547.3243871-1-kai.vehmanen@linux.intel.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda/realtek - Add supported for more Lenovo ALC285 Headset Button
Kailang Yang [Fri, 11 Dec 2020 06:17:09 +0000 (14:17 +0800)]
ALSA: hda/realtek - Add supported for more Lenovo ALC285 Headset Button

Add supported for more Lenovo ALC285 Headset Button.

Signed-off-by: Kailang Yang <kailang@realtek.com>
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/bb1f1da1526d460885aa4257be81eb94@realtek.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda/ca0132 - Remove now unnecessary DSP setup functions.
Connor McAdams [Thu, 10 Dec 2020 16:06:57 +0000 (11:06 -0500)]
ALSA: hda/ca0132 - Remove now unnecessary DSP setup functions.

Now that the DSP's audio configuration is understood, remove previous
hacky methods of trying to properly configure it.

Signed-off-by: Connor McAdams <conmanx360@gmail.com>
Link: https://lore.kernel.org/r/20201210160658.461739-6-conmanx360@gmail.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda/ca0132 - Ensure DSP is properly setup post-firmware download.
Connor McAdams [Thu, 10 Dec 2020 16:06:56 +0000 (11:06 -0500)]
ALSA: hda/ca0132 - Ensure DSP is properly setup post-firmware download.

Make sure that the DSP has no DMA channels allocated once the firmware
is downloaded, and that the default audio streams in use by the DSP are
setup in the correct order.

Signed-off-by: Connor McAdams <conmanx360@gmail.com>
Link: https://lore.kernel.org/r/20201210160658.461739-5-conmanx360@gmail.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda/ca0132 - Add 8051 exram helper functions.
Connor McAdams [Thu, 10 Dec 2020 16:06:55 +0000 (11:06 -0500)]
ALSA: hda/ca0132 - Add 8051 exram helper functions.

Add functions for both reading and writing to the 8051's exram. Also,
add a little bit of documentation on how the addresses are segmented.

Signed-off-by: Connor McAdams <conmanx360@gmail.com>
Link: https://lore.kernel.org/r/20201210160658.461739-4-conmanx360@gmail.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda/ca0132 - Add stream port remapping function.
Connor McAdams [Thu, 10 Dec 2020 16:06:54 +0000 (11:06 -0500)]
ALSA: hda/ca0132 - Add stream port remapping function.

Add function for remapping a ChipIO stream's ports. Also include some
documentation as to how this works.

Signed-off-by: Connor McAdams <conmanx360@gmail.com>
Link: https://lore.kernel.org/r/20201210160658.461739-3-conmanx360@gmail.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda/ca0132 - Reset codec upon initialization.
Connor McAdams [Thu, 10 Dec 2020 16:06:53 +0000 (11:06 -0500)]
ALSA: hda/ca0132 - Reset codec upon initialization.

Reset the codec upon initialization to clear out anything that may have
been setup on a previous boot into Windows, or in case of an improper
shutdown.

Signed-off-by: Connor McAdams <conmanx360@gmail.com>
Link: https://lore.kernel.org/r/20201210160658.461739-2-conmanx360@gmail.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda/hdmi: fix silent stream for first playback to DP
Kai Vehmanen [Thu, 10 Dec 2020 17:44:45 +0000 (19:44 +0200)]
ALSA: hda/hdmi: fix silent stream for first playback to DP

A problem exists in enabling silent stream when connection type is
DisplayPort. Silent stream programming is completed when a new DP
receiver is connected, but infoframe transmission does not actually
start until PCM is opened for the first time. This can result in audible
gap of multiple seconds. This only affects the first PCM open.

Fix the issue by properly assigning a converter to the silent stream,
and modifying the required stream ID programming sequence.

This change only affects Intel display audio codecs.

BugLink: https://github.com/thesofproject/linux/issues/2468
Fixes: 951894cf30f4 ("ALSA: hda/hdmi: Add Intel silent stream support")
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Link: https://lore.kernel.org/r/20201210174445.3134104-1-kai.vehmanen@linux.intel.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda/ca0132 - Change Input Source enum strings.
Connor McAdams [Thu, 10 Dec 2020 17:35:49 +0000 (12:35 -0500)]
ALSA: hda/ca0132 - Change Input Source enum strings.

Change the Input Source enumerated control's strings to make it play
nice with pulseaudio.

Fixes: 7cb9d94c05de9 ("ALSA: hda/ca0132: add alt_select_in/out for R3Di + SBZ")
Cc: <stable@kernel.org>
Signed-off-by: Connor McAdams <conmanx360@gmail.com>
Link: https://lore.kernel.org/r/20201208195223.424753-2-conmanx360@gmail.com
Link: https://lore.kernel.org/r/20201210173550.2968-2-conmanx360@gmail.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda/ca0132 - Fix AE-5 rear headphone pincfg.
Connor McAdams [Thu, 10 Dec 2020 17:35:48 +0000 (12:35 -0500)]
ALSA: hda/ca0132 - Fix AE-5 rear headphone pincfg.

The Windows driver sets the pincfg for the AE-5's rear-headphone to
report as a microphone. This causes issues with Pulseaudio mistakenly
believing there is no headphone plugged in. In Linux, we should instead
set it to be a headphone.

Fixes: a6b0961b39896 ("ALSA: hda/ca0132 - fix AE-5 pincfg")
Cc: <stable@kernel.org>
Signed-off-by: Connor McAdams <conmanx360@gmail.com>
Link: https://lore.kernel.org/r/20201208195223.424753-1-conmanx360@gmail.com
Link: https://lore.kernel.org/r/20201210173550.2968-1-conmanx360@gmail.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Add implicit fb support for Steinberg UR22
Takashi Iwai [Wed, 9 Dec 2020 16:18:35 +0000 (17:18 +0100)]
ALSA: usb-audio: Add implicit fb support for Steinberg UR22

Steinberg UR22 (with USB ID 0499:1509) requires the implicit feedback
for the proper playback, otherwise it causes occasional cracks.
This patch adds the corresponding the quirk table entry with the
recently added generic implicit fb support.

Reported-and-tested-by: Kilian <meschi@posteo.de>
Link: https://lore.kernel.org/r/20201209161835.13625-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda: Fix regressions on clear and reconfig sysfs
Takashi Iwai [Wed, 9 Dec 2020 15:01:19 +0000 (16:01 +0100)]
ALSA: hda: Fix regressions on clear and reconfig sysfs

It seems that the HD-audio clear and reconfig sysfs don't work any
longer after the recent driver core change.  There are multiple issues
around that: the linked list corruption and the dead device handling.
The former issue is fixed by another patch for the driver core itself,
while the latter patch needs to be addressed in HD-audio side.

This patch corresponds to the latter, it recovers those broken
functions by replacing the device detach and attach actions with the
standard core API functions, which are almost equivalent with unbind
and bind actions.

Fixes: 654888327e9f ("driver core: Avoid binding drivers to dead devices")
Cc: <stable@vger.kernel.org>
BugLink: https://bugzilla.kernel.org/show_bug.cgi?id=209207
Link: https://lore.kernel.org/r/20201209150119.7705-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda/realtek - Enable headset mic of ASUS Q524UQK with ALC255
Chris Chiu [Wed, 9 Dec 2020 04:57:30 +0000 (12:57 +0800)]
ALSA: hda/realtek - Enable headset mic of ASUS Q524UQK with ALC255

The ASUS laptop Q524UQK with ALC255 codec can't detect the headset
microphone until ALC255_FIXUP_ASUS_MIC_NO_PRESENCE quirk applied.

Signed-off-by: Chris Chiu <chiu@endlessos.org>
Signed-off-by: Jian-Hong Pan <jhp@endlessos.org>
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/20201209045730.9972-1-chiu@endlessos.org
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda/proc - print DP-MST connections
Kai Vehmanen [Tue, 8 Dec 2020 18:57:36 +0000 (20:57 +0200)]
ALSA: hda/proc - print DP-MST connections

To help in debugging issues with DisplayPort Multi-Stream Transport (aka
DP-MST) support, print information of active connections for each device
of a display audio pin widget.

Example output with the patch with two monitors connected to a DP-MST hub:

Devices: 4
     Dev 00: PD = 0, ELDV = 0, IA = 0, Connections [ 0x03* 0x05 0x07 0x09 ]
     Dev 01: PD = 1, ELDV = 1, IA = 0, Connections [ 0x03* 0x05 0x07 0x09 ]
    *Dev 02: PD = 1, ELDV = 1, IA = 0, Connections [ 0x03 0x05* 0x07 0x09 ]
     Dev 03: PD = 0, ELDV = 0, IA = 0, Connections [ 0x03* 0x05 0x07 0x09 ]
Connection: 4
     0x03 0x05* 0x07 0x09

Format of existing "Connection:" entry is left intact to keep
compatibility.

Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Link: https://lore.kernel.org/r/20201208185736.2877541-1-kai.vehmanen@linux.intel.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda/realtek - Enable headset mic of ASUS X430UN with ALC256
Chris Chiu [Mon, 7 Dec 2020 07:27:55 +0000 (15:27 +0800)]
ALSA: hda/realtek - Enable headset mic of ASUS X430UN with ALC256

The ASUS laptop X430UN with ALC256 can't detect the headset microphone
until ALC256_FIXUP_ASUS_MIC_NO_PRESENCE quirk applied.

Signed-off-by: Chris Chiu <chiu@endlessos.org>
Signed-off-by: Jian-Hong Pan <jhp@endlessos.org>
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/20201207072755.16210-1-chiu@endlessos.org
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: rawmidi: Access runtime->avail always in spinlock
Takashi Iwai [Sun, 6 Dec 2020 08:35:27 +0000 (09:35 +0100)]
ALSA: rawmidi: Access runtime->avail always in spinlock

The runtime->avail field may be accessed concurrently while some
places refer to it without taking the runtime->lock spinlock, as
detected by KCSAN.  Usually this isn't a big problem, but for
consistency and safety, we should take the spinlock at each place
referencing this field.

Reported-by: syzbot+a23a6f1215c84756577c@syzkaller.appspotmail.com
Reported-by: syzbot+3d367d1df1d2b67f5c19@syzkaller.appspotmail.com
Link: https://lore.kernel.org/r/20201206083527.21163-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: seq: Use bool for snd_seq_queue internal flags
Takashi Iwai [Sun, 6 Dec 2020 08:34:56 +0000 (09:34 +0100)]
ALSA: seq: Use bool for snd_seq_queue internal flags

The snd_seq_queue struct contains various flags in the bit fields.
Those are categorized to two different use cases, both of which are
protected by different spinlocks.  That implies that there are still
potential risks of the bad operations for bit fields by concurrent
accesses.

For addressing the problem, this patch rearranges those flags to be
a standard bool instead of a bit field.

Reported-by: syzbot+63cbe31877bb80ef58f5@syzkaller.appspotmail.com
Link: https://lore.kernel.org/r/20201206083456.21110-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda/realtek: make bass spk volume adjustable on a yoga laptop
Hui Wang [Sat, 5 Dec 2020 05:11:30 +0000 (13:11 +0800)]
ALSA: hda/realtek: make bass spk volume adjustable on a yoga laptop

This change could fix 2 issues on this machine:
 - the bass speaker's output volume can't be adjusted, that is because
   the bass speaker is routed to the DAC (Nid 0x6) which has no volume
   control.
 - after plugging a headset with vol+, vol- and pause buttons on it,
   press those buttons, nothing happens, this means those buttons
   don't work at all. This machine has alc287 codec, need to add the
   codec id to the disable/enable_headset_jack_key(), then the headset
   button could work.

The quirk of ALC285_FIXUP_THINKPAD_HEADSET_JACK could fix both of these
2 issues.

Cc: <stable@vger.kernel.org>
Signed-off-by: Hui Wang <hui.wang@canonical.com>
Link: https://lore.kernel.org/r/20201205051130.8122-1-hui.wang@canonical.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Add quirk for Pioneer DJ DDJ-SR2
Takashi Iwai [Mon, 30 Nov 2020 08:37:14 +0000 (09:37 +0100)]
ALSA: usb-audio: Add quirk for Pioneer DJ DDJ-SR2

Yet another quirk for Pioneer DJ DDJ-SR2, which is quite similar like
other DJ DDJ models but with slightly different EPs or channels.

Reported-by: Geraldo <geraldogabriel@gmail.com>
Link: https://lore.kernel.org/r/20201130083714.10640-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda/realtek: Add mute LED quirk to yet another HP x360 model
Takashi Iwai [Sat, 28 Nov 2020 09:00:15 +0000 (10:00 +0100)]
ALSA: hda/realtek: Add mute LED quirk to yet another HP x360 model

HP Spectre x360 Convertible 15" version (SSID 103c:827f) needs the
same quirk to make the mute LED working like other models.
  System Information
    Manufacturer: HP
    Product Name: HP Spectre x360 Convertible 15-bl1XX

  Sound Codec:
    Codec: Realtek ALC295
    Vendor Id: 0x10ec0295
    Subsystem Id: 0x103c827f
    Revision Id: 0x100002

Reported-by: <christoph.plattner@gmx.at>
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/20201128090015.7743-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: remove trailing semicolon in macro definition
Tom Rix [Fri, 27 Nov 2020 19:23:12 +0000 (11:23 -0800)]
ALSA: remove trailing semicolon in macro definition

The macro use will already have a semicolon.

Signed-off-by: Tom Rix <trix@redhat.com>
Link: https://lore.kernel.org/r/20201127192312.2861127-1-trix@redhat.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: ppc: remove redundant checks in PS3 driver probe
Leonard Goehrs [Fri, 27 Nov 2020 15:22:59 +0000 (16:22 +0100)]
ALSA: ppc: remove redundant checks in PS3 driver probe

The check for the FW_FEATURE_PS3_LV1 firmware feature is already performed
in ps3_system_bus_init() before registering the driver. So if the probe
function is actually used, this feature is already known to be available.

The check for the match id is also superfluous; the condition is always
true because the bus' match function (ps3_system_bus_match()) only
considers this driver for devices having:
dev->match_id == snd_ps3_bus_driver_info.match_id.

Suggested-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Leonard Goehrs <l.goehrs@pengutronix.de>
Reviewed-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Tested-by: Geoff Levand <geoff@infradead.org>
Link: https://lore.kernel.org/r/20201127152259.1470079-1-l.goehrs@pengutronix.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: ppc: drop if block with always false condition
Uwe Kleine-König [Thu, 26 Nov 2020 16:59:49 +0000 (17:59 +0100)]
ALSA: ppc: drop if block with always false condition

The remove callback is only called for devices that were probed
successfully before. As the matching probe function cannot complete
without error if dev->match_id != PS3_MATCH_ID_SOUND, we don't have to
check this here.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Geert Uytterhoeven <geert@linux-m68k.org>
Tested-by: Geoff Levand <geoff@infradead.org>
Link: https://lore.kernel.org/r/20201126165950.2554997-1-u.kleine-koenig@pengutronix.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Add support for Pioneer DJ DDJ-RR controller
Daniel Martin [Sat, 28 Nov 2020 08:40:35 +0000 (09:40 +0100)]
ALSA: usb-audio: Add support for Pioneer DJ DDJ-RR controller

This patch allows the Pioneer DJ DDJ-RR to be seen as a USB audio
device under Linux and therefore usable in such applications as
Mixxx.

Tested Master Audio out, headphones (both output jacks) and microphone
input. All work perfectly.

Signed-off-by: Daniel Martin <dmanlfc@gmail.com>
Link: https://lore.kernel.org/r/20201128084035.2958-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda/ca0132: Move unsol callback setups to parser
Takashi Iwai [Wed, 30 Sep 2020 11:30:08 +0000 (13:30 +0200)]
ALSA: hda/ca0132: Move unsol callback setups to parser

The setup of unsolicited event callbacks should be done only once at
the parser phase, not in the init phase that is called multiple times
at each resume.

This patch moves the unsol setup code in ca0132 codec driver to the
more appropriate place.

Reported-by: Hui Wang <hui.wang@canonical.com>
Link: https://lore.kernel.org/r/20200930113008.9307-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda/realtek: Fix bass speaker DAC assignment on Asus Zephyrus G14
Takashi Iwai [Fri, 27 Nov 2020 14:11:04 +0000 (15:11 +0100)]
ALSA: hda/realtek: Fix bass speaker DAC assignment on Asus Zephyrus G14

ASUS Zephyrus G14 has two speaker pins, and the auto-parser tries to
assign an individual DAC to each pin as much as possible.
Unfortunately the third DAC has no volume control unlike the two DACs,
and this resulted in the inconsistent speaker volumes.

As a workaround, wire both speaker pins to the same DAC by modifying
the existing quirk (ALC289_FIXUP_ASUS_GA401) applied to this device.
Since this quirk entry is chained by another, we need to avoid
applying the DAC assignment change for it.  Luckily, there is another
quirk entry (ALC289_FIXUP_ASUS_GA502) doing the very same thing, so we
can chain to the GA502 quirk instead.

Note that this patch uses a new flag of the generic parser,
obey_preferred_dacs, for enforcing the DACs.

BugLink: https://bugzilla.kernel.org/show_bug.cgi?id=210359
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/20201127141104.11041-2-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda/generic: Add option to enforce preferred_dacs pairs
Takashi Iwai [Fri, 27 Nov 2020 14:11:03 +0000 (15:11 +0100)]
ALSA: hda/generic: Add option to enforce preferred_dacs pairs

The generic parser accepts the preferred_dacs[] pairs as a hint for
assigning a DAC to each pin, but this hint doesn't work always
effectively.  Currently it's merely a secondary choice after the trial
with the path index failed.  This made sometimes it difficult to
assign DACs without mimicking the connection list and/or the badness
table.

This patch adds a new flag, obey_preferred_dacs, that changes the
behavior of the parser.  As its name stands, the parser obeys the
given preferred_dacs[] pairs by skipping the path index matching and
giving a high penalty if no DAC is assigned by the pairs.  This mode
will help for assigning the fixed DACs forcibly from the codec
driver.

Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/20201127141104.11041-1-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: compress: allow pause and resume during draining
Gyeongtaek Lee [Thu, 26 Nov 2020 12:34:52 +0000 (21:34 +0900)]
ALSA: compress: allow pause and resume during draining

With a stream with low bitrate, user can't pause or resume the stream
near the end of the stream because current ALSA doesn't allow it.
If the stream has very low bitrate enough to store whole stream into
the buffer, user can't do anything except stop the stream and then
restart it from the first because most of applications call draining
after sending last frame to the kernel.
If pause, resume are allowed during draining, user experience can be
enhanced.
To prevent malfunction in HW drivers which don't support pause
during draining, pause during draining will only work if HW driver
enable this feature explicitly by calling
snd_compr_use_pause_in_draining().

Signed-off-by: Gyeongtaek Lee <gt82.lee@samsung.com>
Acked-by: Vinod Koul <vkoul@kernel.org>
Link: https://lore.kernel.org/r/000101d6c3f0$89b312b0$9d193810$@samsung.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoMerge tag 'asoc-fix-v5.10-rc5' of https://git.kernel.org/pub/scm/linux/kernel/git...
Takashi Iwai [Fri, 27 Nov 2020 14:02:30 +0000 (15:02 +0100)]
Merge tag 'asoc-fix-v5.10-rc5' of https://git./linux/kernel/git/broonie/sound into for-linus

ASoC: Fixes for v5.10

A small set of driver specific fixes, plus a new platform quirk from
Hans.

3 years agoALSA: usb-audio: US16x08: fix value count for level meters
Hector Martin [Fri, 27 Nov 2020 13:26:35 +0000 (22:26 +0900)]
ALSA: usb-audio: US16x08: fix value count for level meters

The level meter control returns 34 integers of info. This fixes:

snd-usb-audio 3-1:1.0: control 2:0:0:Level Meter:0: access overflow

Fixes: d2bb390a2081 ("ALSA: usb-audio: Tascam US-16x08 DSP mixer quirk")
Cc: stable@vger.kernel.org
Signed-off-by: Hector Martin <marcan@marcan.st>
Link: https://lore.kernel.org/r/20201127132635.18947-1-marcan@marcan.st
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: hda/realtek - Add new codec supported for ALC897
Kailang Yang [Fri, 27 Nov 2020 06:39:23 +0000 (14:39 +0800)]
ALSA: hda/realtek - Add new codec supported for ALC897

Enable new codec supported for ALC897.

Signed-off-by: Kailang Yang <kailang@realtek.com>
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/3b00520f304842aab8291eb8d9191bd8@realtek.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoASoC: rt5682: change SAR voltage threshold
Shuming Fan [Thu, 26 Nov 2020 09:27:59 +0000 (17:27 +0800)]
ASoC: rt5682: change SAR voltage threshold

To fix errors in some 4 poles headset detection cases,
this patch adjusts the voltage threshold for mic detection.

Signed-off-by: Shuming Fan <shumingf@realtek.com>
Link: https://lore.kernel.org/r/20201126092759.9427-1-shumingf@realtek.com
Signed-off-by: Mark Brown <broonie@kernel.org>
3 years agoASoC: wm_adsp: fix error return code in wm_adsp_load()
Luo Meng [Mon, 23 Nov 2020 13:38:39 +0000 (21:38 +0800)]
ASoC: wm_adsp: fix error return code in wm_adsp_load()

Fix to return a negative error code from the error handling case
instead of 0 in function wm_adsp_load(), as done elsewhere in this
function.

Fixes: 170b1e123f38 ("ASoC: wm_adsp: Add support for new Halo core DSPs")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Luo Meng <luomeng12@huawei.com>
Acked-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20201123133839.4073787-1-luomeng12@huawei.com
Signed-off-by: Mark Brown <broonie@kernel.org>
3 years agoALSA: hda/realtek: Enable headset of ASUS UX482EG & B9400CEA with ALC294
Jian-Hong Pan [Tue, 24 Nov 2020 09:20:25 +0000 (17:20 +0800)]
ALSA: hda/realtek: Enable headset of ASUS UX482EG & B9400CEA with ALC294

Some laptops like ASUS UX482EG & B9400CEA's headset audio does not work
until the quirk ALC294_FIXUP_ASUS_HPE is applied.

Signed-off-by: Jian-Hong Pan <jhp@endlessos.org>
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/20201124092024.179540-1-jhp@endlessos.org
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoMerge branch 'topic/usb-audio-refactoring' into for-next
Takashi Iwai [Tue, 24 Nov 2020 08:12:37 +0000 (09:12 +0100)]
Merge branch 'topic/usb-audio-refactoring' into for-next

Pull the USB audio improvement

Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoASoC: qcom: Fix enabling BCLK and LRCLK in LPAIF invalid state
Srinivasa Rao Mandadapu [Mon, 23 Nov 2020 16:17:53 +0000 (21:47 +0530)]
ASoC: qcom: Fix enabling BCLK and LRCLK in LPAIF invalid state

Fix enabling BCLK and LRCLK only when LPAIF is invalid state and
bit clock in enable state.
In device suspend/resume scenario LPAIF is going to reset state.
which is causing LRCLK disable and BCLK enable.
Avoid such inconsitency by removing unnecessary cpu dai prepare API,
which is doing LRCLK enable, and by maintaining BLCK  state information.

Fixes: 7e6799d8f87d ("ASoC: qcom: lpass-cpu: Enable MI2S BCLK and LRCLK together")
Signed-off-by: V Sujith Kumar Reddy <vsujithk@codeaurora.org>
Signed-off-by: Srinivasa Rao Mandadapu <srivasam@codeaurora.org>
Link: https://lore.kernel.org/r/1606148273-17325-1-git-send-email-srivasam@codeaurora.org
Signed-off-by: Mark Brown <broonie@kernel.org>
3 years agoALSA: usb-audio: Fix MOTU M-Series quirks
Dylan Robinson [Mon, 23 Nov 2020 08:53:47 +0000 (09:53 +0100)]
ALSA: usb-audio: Fix MOTU M-Series quirks

Now that the usb audio driver correctly finds implicit feedback endpoints,
the implicit feedback quirk for the MOTU M-Series is no longer required.

This also removes some unnecessary vendor specific messages from the MOTU
M-Series boot quirk. The removed vendor specific messages turned on vendor
specific interrupts to the host every 32 samples. The only thing the boot
quirk needs to do is wait for 2 seconds.

Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Signed-off-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-42-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Fix quirks for other BOSS devices
Takashi Iwai [Mon, 23 Nov 2020 08:53:46 +0000 (09:53 +0100)]
ALSA: usb-audio: Fix quirks for other BOSS devices

A few other BOSS devices (BR-80, GT-100v2, Katana) seem requiring the
same quirk as BOSS GT-001, i.e. no implicit feedback for playback but
tying with capture.  Add and correct the corresponding quirk table
entries for them.

Reported-and-tested-by: Keith Milner <kamilner@superlative.org>
Link: https://lore.kernel.org/r/20201123085347.19667-41-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Add implicit_fb module option
Takashi Iwai [Mon, 23 Nov 2020 08:53:45 +0000 (09:53 +0100)]
ALSA: usb-audio: Add implicit_fb module option

A new module option, implicit_fb, is added to specify the driver
looking for the implicit feedback sync.  This can be useful for a
device that could be working better in the implicit feed back mode and
user wants to test it quickly.  When this works, we can add the quirk
entry easily.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-40-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Add generic implicit fb parsing
Takashi Iwai [Mon, 23 Nov 2020 08:53:44 +0000 (09:53 +0100)]
ALSA: usb-audio: Add generic implicit fb parsing

This patch extends the implicit feedback mode parser code to check the
description more generically, so that the quirk entries can be added
without the explicit EP and interface numbers.  The search is done for
the next and the previous interface of the given altset, and if both
entries are ASYNC mode and the direction matches, it just takes as the
sync endpoint.  The generic parser is applicable only for the playback
stream.

As of now, only a few M-Audio devices have been converted to use this
mode.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-39-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Factor out the implicit feedback quirk code
Takashi Iwai [Mon, 23 Nov 2020 08:53:43 +0000 (09:53 +0100)]
ALSA: usb-audio: Factor out the implicit feedback quirk code

The code dealing with the implicit feedback mode grew recently, and
it's becoming messy.  As we receive more and more devices that need
the similar handling, it's better to be processed through a table
instead of the open code.

This patch moves the code that is relevant with parsing the implicit
feedback mode and some helpers into another file, implicit.c.  The
detection and the setup of the implicit feedback sync EPs are
rewritten to use the ID/class matching table instead.

There should be no functional changes.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-38-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Quirk for BOSS GT-001
Takashi Iwai [Mon, 23 Nov 2020 08:53:42 +0000 (09:53 +0100)]
ALSA: usb-audio: Quirk for BOSS GT-001

The capture stream of BOSS GT-001 seems always requiring to be tied
with the playback stream.  OTOH, the playback stream of this device
doesn't seem working in the implicit fb mode, per se, since the
playback must be running before the capture stream.

This patch tries to address the points above:
- Avoid the implicit fb mode for the playback
- Set up a fake sync EP for the capture stream with the hard-coded
  playback stream using the implicit fb mode

Reported-by: Keith Milner <kamilner@superlative.org>
Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-37-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Show sync endpoint information in proc outputs
Takashi Iwai [Mon, 23 Nov 2020 08:53:41 +0000 (09:53 +0100)]
ALSA: usb-audio: Show sync endpoint information in proc outputs

Now the sync endpoints have been parsed at the beginning and won't be
changed dynamically, let's show them in the proc outputs for helping
debugging.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-36-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Use unsigned char for iface and altsettings fields
Takashi Iwai [Mon, 23 Nov 2020 08:53:40 +0000 (09:53 +0100)]
ALSA: usb-audio: Use unsigned char for iface and altsettings fields

Just for consistency, use unsigned char for iface and altsetting in
allover places.  Also rearrange the field positions of
snd_usb_endpiont and tidy up with some comments.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-35-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Replace slave/master terms
Takashi Iwai [Mon, 23 Nov 2020 08:53:39 +0000 (09:53 +0100)]
ALSA: usb-audio: Replace slave/master terms

Follow the inclusive terminology, just replace sync_master/sync_slave
with sync_source/sync_sink.  It's also a bit clearer from its meaning,
too.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-34-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Simplify rate_min/max and rates set up
Takashi Iwai [Mon, 23 Nov 2020 08:53:38 +0000 (09:53 +0100)]
ALSA: usb-audio: Simplify rate_min/max and rates set up

There are multiple places in format.c performing the similar code for
setting the rate_min, rate_max and rates fields.  This patch unifies
those in a helper function and calls it at the end of the parser phase
so that all rate_table entries have been already determined.

No functional changes, just a minor code refactoring.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-33-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Unify the code for the next packet size calculation
Takashi Iwai [Mon, 23 Nov 2020 08:53:37 +0000 (09:53 +0100)]
ALSA: usb-audio: Unify the code for the next packet size calculation

There are two places calculating the next packet size for the playback
stream in the exactly same way.  Provide the single helper for this
purpose and use it from both places gracefully.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-32-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Drop unneeded snd_usb_substream fields
Takashi Iwai [Mon, 23 Nov 2020 08:53:36 +0000 (09:53 +0100)]
ALSA: usb-audio: Drop unneeded snd_usb_substream fields

Some fields like interface and alt_idx in snd_usb_substream are mostly
useless now as they can be referred via either cur_audiofmt or
data_endpoint assigned to the substream.  Drop those, and also assure
the concurrency about the access of cur_audiofmt field.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-31-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Refactoring endpoint URB deactivation
Takashi Iwai [Mon, 23 Nov 2020 08:53:35 +0000 (09:53 +0100)]
ALSA: usb-audio: Refactoring endpoint URB deactivation

Minor code refactoring to consolidate the URB deactivation code in
endpoint.c.  A slight behavior change is that the error handling in
snd_usb_endpoint_start() leaves EP_FLAG_STOPPING now.  This should be
synced with the later PCM sync_stop callback.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-30-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Use atomic_t for endpoint use_count
Takashi Iwai [Mon, 23 Nov 2020 08:53:34 +0000 (09:53 +0100)]
ALSA: usb-audio: Use atomic_t for endpoint use_count

The endpoint objects may be started/stopped concurrently by different
substreams in the case of implicit feedback mode, while the current
code handles the reference counter without any protection.

This patch changes the refcount to atomic_t for avoiding the
inconsistency.  We need no reference_t here as the refcount goes only
up to 2.

Also the name "use_count" is renamed to "running" since this is about
actually the running status, not the open refcount.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-29-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Constify audioformat pointer references
Takashi Iwai [Mon, 23 Nov 2020 08:53:33 +0000 (09:53 +0100)]
ALSA: usb-audio: Constify audioformat pointer references

The audioformat is referred in many places but most of usages are
read-only.  Let's add const prefix in the possible places.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-28-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Fix possible stall of implicit fb packet ring-buffer
Takashi Iwai [Mon, 23 Nov 2020 08:53:32 +0000 (09:53 +0100)]
ALSA: usb-audio: Fix possible stall of implicit fb packet ring-buffer

The implicit feedback mode uses a ring buffer for storing the received
packet sizes from the feedback source, and the code has a slight flaw;
when a playback stream stalls by some reason and the URBs aren't
processed, the next_packet FIFO might become empty, but the driver
can't distinguish whether it's empty or full because it's managed with
read_poss and write_pos.

This patch addresses those by changing the next_packet array
management.  Instead of keeping read and write positions, now the head
position and the queued amount are kept.  It's easier to understand
about the emptiness.  Also, the URB active flag is now cleared before
calling queue_pending_output_urbs() for avoiding (theoretically)
possible inconsistency.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-27-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Refactor endpoint management
Takashi Iwai [Mon, 23 Nov 2020 08:53:31 +0000 (09:53 +0100)]
ALSA: usb-audio: Refactor endpoint management

This is an intensive surgery for the endpoint and stream management
for achieving more robust and clean code.

The goals of this patch are:
- More clear endpoint resource changes
- The interface altsetting control in a single place
Below are brief description of the whole changes.

First off, most of the endpoint operations are moved into endpoint.c,
so that the snd_usb_endpoint object is only referred in other places.
The endpoint object is acquired and released via the new functions
snd_usb_endpoint_open() and snd_usb_endpoint_close() that are called
at PCM hw_params and hw_free callbacks, respectively.  Those are
ref-counted and EPs can manage the multiple opens.

The open callback receives the audioformat and hw_params arguments,
and those are used for initializing the EP parameters; especially the
endpoint, interface and altset numbers are read from there, as well as
the PCM parameters like the format, rate and channels.  Those are
stored in snd_usb_endpoint object.  If it's the secondary open, the
function checks whether the given parameters are compatible with the
already opened EP setup, too.

The coupling with a sync EP (including an implicit feedback sync) is
done by the sole snd_usb_endpoint_set_sync() call.

The configuration of each endpoint is done in a single shot via
snd_usb_endpoint_configure() call.  This is the place where most of
PCM configurations are done.  A few flags and special handling in the
snd_usb_substream are dropped along with this change.

A significant difference wrt the configuration from the previous code
is the order of USB host interface setups.  Now the interface is
always disabled at beginning and (re-)enabled at the last step of
snd_usb_endpoint_configure(), in order to be compliant with the
standard UAC2/3.  For UAC1, the interface is set before the parameter
setups since there seem devices that require it (e.g. Yamaha THR10),
just like how it was done in the previous driver code.

The start/stop are almost same as before, also single-shots.  The URB
callbacks need to be set via snd_usb_endpoint_set_callback() like the
previous code at the trigger phase, too.

Finally, the flag for the re-setup is set at the device suspend
through the full EP list, instead of PCM trigger.  This catches the
overlooked cases where the PCM hasn't been running yet but the device
needs the full setup after resume.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-26-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Fix EP matching for continuous rates
Takashi Iwai [Mon, 23 Nov 2020 08:53:30 +0000 (09:53 +0100)]
ALSA: usb-audio: Fix EP matching for continuous rates

The function to evaluate the match of the parameters with an EP
assumes only the discrete rate tables and doesn't handle the
continuous rates properly.

This patch fixes match_endpoint_audioformats() to handle the
continuous rates.  Also the almost useless debug prints there are
dropped.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-25-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Always set up the parameters after resume
Takashi Iwai [Mon, 23 Nov 2020 08:53:29 +0000 (09:53 +0100)]
ALSA: usb-audio: Always set up the parameters after resume

The commit 92adc96f8eec ("ALSA: usb-audio: set the interface format
after resume on Dell WD19") introduced the workaround for the broken
setup after the resume specifically on a Dell dock model.  However,
the full setup should have been performed after the resume on all
devices, as we can't guarantee the same state.  So this patch removes
the conditional check and applies the workaround always.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-24-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Set callbacks via snd_usb_endpoint_set_callback()
Takashi Iwai [Mon, 23 Nov 2020 08:53:28 +0000 (09:53 +0100)]
ALSA: usb-audio: Set callbacks via snd_usb_endpoint_set_callback()

The prepare_data_urb and retire_data_urb fields of the endpoint object
are set dynamically at PCM trigger start/stop.  Those are evaluated in
the endpoint handler, but there can be a race, especially if two
different PCM substreams are handling the same endpoint for the
implicit feedback case.  Also, the data_subs field of the endpoint is
set and accessed dynamically, too, which has the same risk.

As a slight improvement for the concurrency, this patch introduces the
function to set the callbacks and the data in a shot with the memory
barrier.  In the reader side, it's also fetched with the memory
barrier.

There is still a room of race if prepare and retire callbacks are set
during executing the URB completion.  But such an inconsistency may
happen only for the implicit fb source, i.e. it's only about the
capture stream.  And luckily, the capture stream never sets the
prepare callback, hence the problem doesn't happen practically.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-23-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Stop both endpoints properly at error
Takashi Iwai [Mon, 23 Nov 2020 08:53:27 +0000 (09:53 +0100)]
ALSA: usb-audio: Stop both endpoints properly at error

start_endpoints() may leave the data endpoint running if an error
happens at starting the sync endpoint.  We should stop both streams
properly, instead.

While we're at it, move the debug prints into the endpoint.c that is a
more suitable place.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-22-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Simplify snd_usb_init_pitch() arguments
Takashi Iwai [Mon, 23 Nov 2020 08:53:26 +0000 (09:53 +0100)]
ALSA: usb-audio: Simplify snd_usb_init_pitch() arguments

A preliminary change for the later big changes.  This is a minor code
refactoring to drop the unnecessary arguments that can be retrieved in
a different way.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-21-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Simplify snd_usb_init_sample_rate() arguments
Takashi Iwai [Mon, 23 Nov 2020 08:53:25 +0000 (09:53 +0100)]
ALSA: usb-audio: Simplify snd_usb_init_sample_rate() arguments

A preliminary change for the later big changes.  This is a minor code
refactoring to drop the unnecessary arguments that can be retrieved in
a different way.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-20-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Pass snd_usb_audio object to quirk functions
Takashi Iwai [Mon, 23 Nov 2020 08:53:24 +0000 (09:53 +0100)]
ALSA: usb-audio: Pass snd_usb_audio object to quirk functions

A preliminary patch for the later big change.  Just a minor code
refactoring.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-19-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Don't set altsetting before initializing sample rate
Takashi Iwai [Mon, 23 Nov 2020 08:53:23 +0000 (09:53 +0100)]
ALSA: usb-audio: Don't set altsetting before initializing sample rate

Setting the active altsetting at changing sample rate seems
unrecommended.  The host should deselect the altsetting at first
before that, then select it again.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-18-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Add snd_usb_get_host_interface() helper
Takashi Iwai [Mon, 23 Nov 2020 08:53:22 +0000 (09:53 +0100)]
ALSA: usb-audio: Add snd_usb_get_host_interface() helper

Add a helper function to retrieve the usb_host_interface object from
the given interface and altsetting number pair, which is a commonly
used procedure in the driver code.

No functional changes, just minor code refactoring.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-17-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Drop keep_interface flag again
Takashi Iwai [Mon, 23 Nov 2020 08:53:21 +0000 (09:53 +0100)]
ALSA: usb-audio: Drop keep_interface flag again

This behavior turned out to be invalid from the USB spec POV and
shouldn't be applied.  As it's an optional flag that is set only via
an card control element that must be hardly used, let's drop it
again.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-16-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Create endpoint objects at parsing phase
Takashi Iwai [Mon, 23 Nov 2020 08:53:20 +0000 (09:53 +0100)]
ALSA: usb-audio: Create endpoint objects at parsing phase

Currently snd_usb_endpoint objects are created at first when the
substream is opened and tries to assign the endpoints corresponding to
the matching audioformat.  But since basically the all endpoints have
been already parsed and the information have been obtained, we may
create the endpoint objects statically at the init phase.  It's easier
to manage for the implicit fb case, for example.

This patch changes the endpoint object management and lets the parser
to create the all endpoint objects.

This change shouldn't bring any functional changes.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-15-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Avoid doubly initialization for implicit fb
Takashi Iwai [Mon, 23 Nov 2020 08:53:19 +0000 (09:53 +0100)]
ALSA: usb-audio: Avoid doubly initialization for implicit fb

The implicit feedback mode initializes both the main data stream and
the sync data stream.  When a sync stream was already opened, this
would result in the doubly initialization and might screw up things.

Add the check of already opened sync streams and skip the unnecessary
initialization.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-14-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Drop debug.h
Takashi Iwai [Mon, 23 Nov 2020 08:53:18 +0000 (09:53 +0100)]
ALSA: usb-audio: Drop debug.h

The file debug.h contains a simple macro for debug prints, and it's
used only in two places, the format parser and the hw_params rules.
The former actually should print a more informative message instead,
so the only users are the hw_parmas rules.

This patch moves the contents of debug.h into the hw_params rules
local code and remove the unneeded includes.  Also, the debug print in
the format parser is replaced with the information print with more
useful information, and the raw printk() call is replaced with
pr_debug().

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-13-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Simplify hw_params rules
Takashi Iwai [Mon, 23 Nov 2020 08:53:17 +0000 (09:53 +0100)]
ALSA: usb-audio: Simplify hw_params rules

Several hw_params functions narrows the interval via min/max rule in
the very similar way, so factor out those into a helper function and
use commonly.

No functional changes, just minor code refactoring.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-12-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Add hw constraint for implicit fb sync
Takashi Iwai [Mon, 23 Nov 2020 08:53:16 +0000 (09:53 +0100)]
ALSA: usb-audio: Add hw constraint for implicit fb sync

In the current code, there is no check at the stream open time whether
the endpoint is being already used by others.  In the normal
operations, this shouldn't happen, but in the case of the implicit
feedback mode, it's a common problem with the full duplex operation,
because the capture stream is always opened by the playback stream as
an implicit sync source.

Although we recently introduced the check of such a conflict of
parameters at the PCM hw_params time, it doesn't give any hint at the
hw_params itself and just gives the error.  This isn't quite
comfortable, and it caused problems on many applications.

This patch attempts to make the parameter handling easier by
introducing the strict hw constraint matching with the counterpart
stream that is being used.  That said, when an implicit feedback
playback stream is running before a capture stream is opened, the
capture stream carries the PCM hw-constraint to allow only the same
sample rate, format, periods and period frames as the running playback
stream.  If not opened or there is no conflict of endpoints, the
behavior remains as same as before.

Note that this kind of "weak link" should work for most cases, but
this is no concrete solution; e.g. if an application changes the hw
params multiple times while another stream is opened, this would lead
to inconsistencies.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-11-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Move snd_usb_autoresume() call out of setup_hw_info()
Takashi Iwai [Mon, 23 Nov 2020 08:53:15 +0000 (09:53 +0100)]
ALSA: usb-audio: Move snd_usb_autoresume() call out of setup_hw_info()

This is a preliminary work for the upcoming hw-constraint change for
the implicit feedback mode.

Currently snd_usb_autoresume() is called at the end of
setup_hwinfo().  It's a bit confusing; because of this implicit
refcount usage, the caller side needs to call snd_usb_autosuspend()
later in the error path although it's not seen inside the function.
Instead, it's clearer to call both snd_usb_autoresume() and suspend()
in the very same function.

It's only refactoring and no functional changes.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-10-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Track implicit fb sync endpoint in audioformat list
Takashi Iwai [Mon, 23 Nov 2020 08:53:14 +0000 (09:53 +0100)]
ALSA: usb-audio: Track implicit fb sync endpoint in audioformat list

Instead of parsing and evaluating the sync endpoint and the implicit
feedback mode at each time the audio stream is opened, let's parse it
once at the probe time, as the all needed information can be obtained
statically from the descriptor or from the quirk.

This patch extends audioformat struct to record the sync endpoint,
interface and altsetting as well as the implicit feedback flag, which
are filled at parsing the streams.  Then, set_sync_endpoint() is much
simplified just to follow the already parsed data.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-9-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Improve some debug prints
Takashi Iwai [Mon, 23 Nov 2020 08:53:13 +0000 (09:53 +0100)]
ALSA: usb-audio: Improve some debug prints

There are a few rooms for improvements wrt the debug prints:
- The EP debug print is shown only at starting, not at stopping
- The EP debug print contains useless object addresses
- Some helpers show the urb and the EP object addresses, too

This patch addresses those shortcomings.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-8-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Set and clear sync EP link properly
Takashi Iwai [Mon, 23 Nov 2020 08:53:12 +0000 (09:53 +0100)]
ALSA: usb-audio: Set and clear sync EP link properly

The sync EP setup isn't cleared at stopping the stream but expected to
be cleared at the next stream start.  This may leave the sync link
setup stale and can spoof wrongly when full duplex streams were
running in the implicit fb sync.  Let's initialize them properly at
start and end of the stream.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-7-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Add snd_usb_get_endpoint() helper
Takashi Iwai [Mon, 23 Nov 2020 08:53:11 +0000 (09:53 +0100)]
ALSA: usb-audio: Add snd_usb_get_endpoint() helper

Factor out the code to obtain snd_usb_endpoint object matching with
the given endpoint.  It'll be used in the later patch to add the
implicit feedback hw-constraint.

No functional change by this patch itself.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-6-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Check implicit feedback EP generically for UAC2
Takashi Iwai [Mon, 23 Nov 2020 08:53:10 +0000 (09:53 +0100)]
ALSA: usb-audio: Check implicit feedback EP generically for UAC2

It seems that many UAC2 devices are with the implicit feedback, but
they couldn't be probed properly because the assumption the driver
takes currently isn't applied: they have the single endpoint for both
data and implicit-fb streams, while we checked only the classical sync
endpoints assigned to the next altsetting in the same interface.

This patch extends the search to match with those typical cases where
the implicit fb stream is found in the next interface number.

While we're at it, slightly refactor the code, not returning 0/-ERROR
but use the standard bool to success/failur, which is more intuitive
in this particular case.

Reported-by: Dylan Robinson <dylan_robinson@motu.com>
Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-5-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Check valid altsetting at parsing rates for UAC2/3
Takashi Iwai [Mon, 23 Nov 2020 08:53:09 +0000 (09:53 +0100)]
ALSA: usb-audio: Check valid altsetting at parsing rates for UAC2/3

The current driver code assumes blindly that all found sample rates for
the same endpoint from the UAC2 and UAC3 descriptors can be used no
matter which altsetting, but actually this was wrong: some devices
accept only limited sample rates in each altsetting.  For determining
which altsetting supports which rate, we need to verify each sample rate
and check the validity via UAC2_AS_VAL_ALT_SETTINGS.  This control
reports back the available altsettings as a bitmap.

This patch implements the missing piece above, the verification and
reconstructs the sample rate tables based on the result.

An open question is how to deal with the altsettings that ended up
with no valid sample rates after verification.  At least, there is a
device that showed this problem although the sample rates did work in
the later usage (see bug link).  For now, we accept such an altset as
is, assuming that it's a firmware bug.

Reported-by: Dylan Robinson <dylan_robinson@motu.com>
Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
BugLink: https://bugzilla.suse.com/show_bug.cgi?id=1178203
Link: https://lore.kernel.org/r/20201123085347.19667-4-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Don't call usb_set_interface() at trigger callback
Takashi Iwai [Mon, 23 Nov 2020 08:53:08 +0000 (09:53 +0100)]
ALSA: usb-audio: Don't call usb_set_interface() at trigger callback

The PCM trigger callback is atomic, hence we must not call a function
like usb_set_interface() there.  Calling it from there would lead to a
kernel Oops.

Fix it by moving the usb_set_interface() call to set_sync_endpoint().

Also, apply the snd_usb_set_interface_quirk() for consistency, too.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-3-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoALSA: usb-audio: Handle discrete rates properly in hw constraints
Takashi Iwai [Mon, 23 Nov 2020 08:53:07 +0000 (09:53 +0100)]
ALSA: usb-audio: Handle discrete rates properly in hw constraints

In the current code, when the device provides the discrete sample rate
tables with unusual sample rates, the driver tries to gather the whole
values from the audioformat entries and create a hw-constraint rule to
restrict with this single rate list.  This is rather inefficient and
may overlook the rates that are associated only with the certain
audioformat entries.

This patch improves the hw constraint setup by rewriting the existing
hw_rule_rate().  The discrete sample rates (identified by rate_table
and nr_rates of format entry) are checked in the existing
hw_rule_rate() instead of extra rules; in the case of discrete rates,
the function compares with each rate table entry and calculates the
min/max values from there.  For the contiguous rates, the behavior
doesn't change.

Along with it, snd_usb_pcm_check_knot() and snb_usb_substream
rate_list field become superfluous, thus those are dropped.

Tested-by: Keith Milner <kamilner@superlative.org>
Tested-by: Dylan Robinson <dylan_robinson@motu.com>
Link: https://lore.kernel.org/r/20201123085347.19667-2-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
3 years agoLinux 5.10-rc5 v5.10-rc5
Linus Torvalds [Sun, 22 Nov 2020 23:36:08 +0000 (15:36 -0800)]
Linux 5.10-rc5

3 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid
Linus Torvalds [Sun, 22 Nov 2020 22:36:06 +0000 (14:36 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/hid/hid

Pull HID fixes from Jiri Kosina:

 - Various functionality / regression fixes for Logitech devices from
   Hans de Goede

 - Fix for (recently added) GPIO support in mcp2221 driver from Lars
   Povlsen

 - Power management handling fix/quirk in i2c-hid driver for certain
   BIOSes that have strange aproach to power-cycle from Hans de Goede

 - a few device ID additions and device-specific quirks

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid:
  HID: logitech-dj: Fix Dinovo Mini when paired with a MX5x00 receiver
  HID: logitech-dj: Fix an error in mse_bluetooth_descriptor
  HID: Add Logitech Dinovo Edge battery quirk
  HID: logitech-hidpp: Add HIDPP_CONSUMER_VENDOR_KEYS quirk for the Dinovo Edge
  HID: logitech-dj: Handle quad/bluetooth keyboards with a builtin trackpad
  HID: add HID_QUIRK_INCREMENT_USAGE_ON_DUPLICATE for Gamevice devices
  HID: mcp2221: Fix GPIO output handling
  HID: hid-sensor-hub: Fix issue with devices with no report ID
  HID: i2c-hid: Put ACPI enumerated devices in D3 on shutdown
  HID: add support for Sega Saturn
  HID: cypress: Support Varmilo Keyboards' media hotkeys
  HID: ite: Replace ABS_MISC 120/121 events with touchpad on/off keypresses
  HID: logitech-hidpp: Add PID for MX Anywhere 2
  HID: uclogic: Add ID for Trust Flex Design Tablet

3 years agoMerge tag 'sched-urgent-2020-11-22' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 22 Nov 2020 21:26:07 +0000 (13:26 -0800)]
Merge tag 'sched-urgent-2020-11-22' of git://git./linux/kernel/git/tip/tip

Pull scheduler fixes from Thomas Gleixner:
 "A couple of scheduler fixes:

   - Make the conditional update of the overutilized state work
     correctly by caching the relevant flags state before overwriting
     them and checking them afterwards.

   - Fix a data race in the wakeup path which caused loadavg on ARM64
     platforms to become a random number generator.

   - Fix the ordering of the iowaiter accounting operations so it can't
     be decremented before it is incremented.

   - Fix a bug in the deadline scheduler vs. priority inheritance when a
     non-deadline task A has inherited the parameters of a deadline task
     B and then blocks on a non-deadline task C.

     The second inheritance step used the static deadline parameters of
     task A, which are usually 0, instead of further propagating task
     B's parameters. The zero initialized parameters trigger a bug in
     the deadline scheduler"

* tag 'sched-urgent-2020-11-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched/deadline: Fix priority inheritance with multiple scheduling classes
  sched: Fix rq->nr_iowait ordering
  sched: Fix data-race in wakeup
  sched/fair: Fix overutilized update in enqueue_task_fair()

3 years agoMerge tag 'perf-urgent-2020-11-22' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 22 Nov 2020 21:23:43 +0000 (13:23 -0800)]
Merge tag 'perf-urgent-2020-11-22' of git://git./linux/kernel/git/tip/tip

Pull perf fix from Thomas Gleixner:
 "A single fix for the x86 perf sysfs interfaces which used kobject
  attributes instead of device attributes and therefore making clang's
  control flow integrity checker upset"

* tag 'perf-urgent-2020-11-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf/x86: fix sysfs type mismatches

3 years agoMerge tag 'locking-urgent-2020-11-22' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 22 Nov 2020 21:19:53 +0000 (13:19 -0800)]
Merge tag 'locking-urgent-2020-11-22' of git://git./linux/kernel/git/tip/tip

Pull locking fix from Thomas Gleixner:
 "A single fix for lockdep which makes the recursion protection cover
  graph lock/unlock"

* tag 'locking-urgent-2020-11-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  lockdep: Put graph lock/unlock under lock_recursion protection

3 years agoMerge tag 'efi-urgent-for-v5.10-rc3' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 22 Nov 2020 21:05:48 +0000 (13:05 -0800)]
Merge tag 'efi-urgent-for-v5.10-rc3' of git://git./linux/kernel/git/tip/tip

Pull EFI fixes from Borislav Petkov:
 "Forwarded EFI fixes from Ard Biesheuvel:

   - fix memory leak in efivarfs driver

   - fix HYP mode issue in 32-bit ARM version of the EFI stub when built
     in Thumb2 mode

   - avoid leaking EFI pgd pages on allocation failure"

* tag 'efi-urgent-for-v5.10-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  efi/x86: Free efi_pgd with free_pages()
  efivarfs: fix memory leak in efivarfs_create()
  efi/arm: set HSCTLR Thumb2 bit correctly for HVC calls from HYP

3 years agoMerge tag 'x86_urgent_for_v5.10-rc5' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 22 Nov 2020 20:55:50 +0000 (12:55 -0800)]
Merge tag 'x86_urgent_for_v5.10-rc5' of git://git./linux/kernel/git/tip/tip

Pull x86 fixes from Borislav Petkov:

 - An IOMMU VT-d build fix when CONFIG_PCI_ATS=n along with a revert of
   same because the proper one is going through the IOMMU tree (Thomas
   Gleixner)

 - An Intel microcode loader fix to save the correct microcode patch to
   apply during resume (Chen Yu)

 - A fix to not access user memory of other processes when dumping
   opcode bytes (Thomas Gleixner)

* tag 'x86_urgent_for_v5.10-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  Revert "iommu/vt-d: Take CONFIG_PCI_ATS into account"
  x86/dumpstack: Do not try to access user space code of other tasks
  x86/microcode/intel: Check patch signature before saving microcode for early loading
  iommu/vt-d: Take CONFIG_PCI_ATS into account

3 years agoMerge branch 'akpm' (patches from Andrew)
Linus Torvalds [Sun, 22 Nov 2020 20:14:46 +0000 (12:14 -0800)]
Merge branch 'akpm' (patches from Andrew)

Merge misc fixes from Andrew Morton:
 "8 patches.

  Subsystems affected by this patch series: mm (madvise, pagemap,
  readahead, memcg, userfaultfd), kbuild, and vfs"

* emailed patches from Andrew Morton <akpm@linux-foundation.org>:
  mm: fix madvise WILLNEED performance problem
  libfs: fix error cast of negative value in simple_attr_write()
  mm/userfaultfd: do not access vma->vm_mm after calling handle_userfault()
  mm: memcg/slab: fix root memcg vmstats
  mm: fix readahead_page_batch for retry entries
  mm: fix phys_to_target_node() and memory_add_physaddr_to_nid() exports
  compiler-clang: remove version check for BPF Tracing
  mm/madvise: fix memory leak from process_madvise

3 years agoMerge tag 'staging-5.10-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Sun, 22 Nov 2020 19:58:49 +0000 (11:58 -0800)]
Merge tag 'staging-5.10-rc5' of git://git./linux/kernel/git/gregkh/staging

Pull staging and IIO fixes from Greg KH:
 "Here are some small Staging and IIO driver fixes for 5.10-rc5. They
  include:

   - IIO fixes for reported regressions and problems

   - new device ids for IIO drivers

   - new device id for rtl8723bs driver

   - staging ralink driver Kconfig dependency fix

   - staging mt7621-pci bus resource fix

  All of these have been in linux-next all week with no reported issues"

* tag 'staging-5.10-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging:
  iio: accel: kxcjk1013: Add support for KIOX010A ACPI DSM for setting tablet-mode
  iio: accel: kxcjk1013: Replace is_smo8500_device with an acpi_type enum
  docs: ABI: testing: iio: stm32: remove re-introduced unsupported ABI
  iio: light: fix kconfig dependency bug for VCNL4035
  iio/adc: ingenic: Fix AUX/VBAT readings when touchscreen is used
  iio/adc: ingenic: Fix battery VREF for JZ4770 SoC
  staging: rtl8723bs: Add 024c:0627 to the list of SDIO device-ids
  staging: ralink-gdma: fix kconfig dependency bug for DMA_RALINK
  staging: mt7621-pci: avoid to request pci bus resources
  iio: imu: st_lsm6dsx: set 10ms as min shub slave timeout
  counter/ti-eqep: Fix regmap max_register
  iio: adc: stm32-adc: fix a regression when using dma and irq
  iio: adc: mediatek: fix unset field
  iio: cros_ec: Use default frequencies when EC returns invalid information

3 years agoMerge tag 'tty-5.10-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty
Linus Torvalds [Sun, 22 Nov 2020 19:52:10 +0000 (11:52 -0800)]
Merge tag 'tty-5.10-rc5' of git://git./linux/kernel/git/gregkh/tty

Pull tty fixes from Greg KH:
 "Here are some small tty/serial fixes for 5.10-rc5 that resolve some
  reported issues:

   - speakup crash when telling the kernel to use a device that isn't
     really there

   - imx serial driver fixes for reported problems

   - ar933x_uart driver fix for probe error handling path

  All have been in linux-next for a while with no reported issues"

* tag 'tty-5.10-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty:
  serial: ar933x_uart: disable clk on error handling path in probe
  tty: serial: imx: keep console clocks always on
  speakup: Do not let the line discipline be used several times
  tty: serial: imx: fix potential deadlock

3 years agoMerge tag 'ext4_for_linus_fixes2' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 22 Nov 2020 19:39:32 +0000 (11:39 -0800)]
Merge tag 'ext4_for_linus_fixes2' of git://git./linux/kernel/git/tytso/ext4

Pull ext4 fixes from Ted Ts'o:
 "A final set of miscellaneous bug fixes for ext4"

* tag 'ext4_for_linus_fixes2' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
  ext4: fix bogus warning in ext4_update_dx_flag()
  jbd2: fix kernel-doc markups
  ext4: drop fast_commit from /proc/mounts

3 years agoafs: Fix speculative status fetch going out of order wrt to modifications
David Howells [Sun, 22 Nov 2020 13:13:45 +0000 (13:13 +0000)]
afs: Fix speculative status fetch going out of order wrt to modifications

When doing a lookup in a directory, the afs filesystem uses a bulk
status fetch to speculatively retrieve the statuses of up to 48 other
vnodes found in the same directory and it will then either update extant
inodes or create new ones - effectively doing 'lookup ahead'.

To avoid the possibility of deadlocking itself, however, the filesystem
doesn't lock all of those inodes; rather just the directory inode is
locked (by the VFS).

When the operation completes, afs_inode_init_from_status() or
afs_apply_status() is called, depending on whether the inode already
exists, to commit the new status.

A case exists, however, where the speculative status fetch operation may
straddle a modification operation on one of those vnodes.  What can then
happen is that the speculative bulk status RPC retrieves the old status,
and whilst that is happening, the modification happens - which returns
an updated status, then the modification status is committed, then we
attempt to commit the speculative status.

This results in something like the following being seen in dmesg:

kAFS: vnode modified {100058:861} 8->9 YFS.InlineBulkStatus

showing that for vnode 861 on volume 100058, we saw YFS.InlineBulkStatus
say that the vnode had data version 8 when we'd already recorded version
9 due to a local modification.  This was causing the cache to be
invalidated for that vnode when it shouldn't have been.  If it happens
on a data file, this might lead to local changes being lost.

Fix this by ignoring speculative status updates if the data version
doesn't match the expected value.

Note that it is possible to get a DV regression if a volume gets
restored from a backup - but we should get a callback break in such a
case that should trigger a recheck anyway.  It might be worth checking
the volume creation time in the volsync info and, if a change is
observed in that (as would happen on a restore), invalidate all caches
associated with the volume.

Fixes: 5cf9dd55a0ec ("afs: Prospectively look up extra files when doing a single lookup")
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
3 years agomm: fix madvise WILLNEED performance problem
Matthew Wilcox (Oracle) [Sun, 22 Nov 2020 06:17:22 +0000 (22:17 -0800)]
mm: fix madvise WILLNEED performance problem

The calculation of the end page index was incorrect, leading to a
regression of 70% when running stress-ng.

With this fix, we instead see a performance improvement of 3%.

Fixes: e6e88712e43b ("mm: optimise madvise WILLNEED")
Reported-by: kernel test robot <rong.a.chen@intel.com>
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Tested-by: Xing Zhengjun <zhengjun.xing@linux.intel.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: William Kucharski <william.kucharski@oracle.com>
Cc: Feng Tang <feng.tang@intel.com>
Cc: "Chen, Rong A" <rong.a.chen@intel.com>
Link: https://lkml.kernel.org/r/20201109134851.29692-1-willy@infradead.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
3 years agolibfs: fix error cast of negative value in simple_attr_write()
Yicong Yang [Sun, 22 Nov 2020 06:17:19 +0000 (22:17 -0800)]
libfs: fix error cast of negative value in simple_attr_write()

The attr->set() receive a value of u64, but simple_strtoll() is used for
doing the conversion.  It will lead to the error cast if user inputs a
negative value.

Use kstrtoull() instead of simple_strtoll() to convert a string got from
the user to an unsigned value.  The former will return '-EINVAL' if it
gets a negetive value, but the latter can't handle the situation
correctly.  Make 'val' unsigned long long as what kstrtoull() takes,
this will eliminate the compile warning on no 64-bit architectures.

Fixes: f7b88631a897 ("fs/libfs.c: fix simple_attr_write() on 32bit machines")
Signed-off-by: Yicong Yang <yangyicong@hisilicon.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Link: https://lkml.kernel.org/r/1605341356-11872-1-git-send-email-yangyicong@hisilicon.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
3 years agomm/userfaultfd: do not access vma->vm_mm after calling handle_userfault()
Gerald Schaefer [Sun, 22 Nov 2020 06:17:15 +0000 (22:17 -0800)]
mm/userfaultfd: do not access vma->vm_mm after calling handle_userfault()

Alexander reported a syzkaller / KASAN finding on s390, see below for
complete output.

In do_huge_pmd_anonymous_page(), the pre-allocated pagetable will be
freed in some cases.  In the case of userfaultfd_missing(), this will
happen after calling handle_userfault(), which might have released the
mmap_lock.  Therefore, the following pte_free(vma->vm_mm, pgtable) will
access an unstable vma->vm_mm, which could have been freed or re-used
already.

For all architectures other than s390 this will go w/o any negative
impact, because pte_free() simply frees the page and ignores the
passed-in mm.  The implementation for SPARC32 would also access
mm->page_table_lock for pte_free(), but there is no THP support in
SPARC32, so the buggy code path will not be used there.

For s390, the mm->context.pgtable_list is being used to maintain the 2K
pagetable fragments, and operating on an already freed or even re-used
mm could result in various more or less subtle bugs due to list /
pagetable corruption.

Fix this by calling pte_free() before handle_userfault(), similar to how
it is already done in __do_huge_pmd_anonymous_page() for the WRITE /
non-huge_zero_page case.

Commit 6b251fc96cf2c ("userfaultfd: call handle_userfault() for
userfaultfd_missing() faults") actually introduced both, the
do_huge_pmd_anonymous_page() and also __do_huge_pmd_anonymous_page()
changes wrt to calling handle_userfault(), but only in the latter case
it put the pte_free() before calling handle_userfault().

  BUG: KASAN: use-after-free in do_huge_pmd_anonymous_page+0xcda/0xd90 mm/huge_memory.c:744
  Read of size 8 at addr 00000000962d6988 by task syz-executor.0/9334

  CPU: 1 PID: 9334 Comm: syz-executor.0 Not tainted 5.10.0-rc1-syzkaller-07083-g4c9720875573 #0
  Hardware name: IBM 3906 M04 701 (KVM/Linux)
  Call Trace:
    do_huge_pmd_anonymous_page+0xcda/0xd90 mm/huge_memory.c:744
    create_huge_pmd mm/memory.c:4256 [inline]
    __handle_mm_fault+0xe6e/0x1068 mm/memory.c:4480
    handle_mm_fault+0x288/0x748 mm/memory.c:4607
    do_exception+0x394/0xae0 arch/s390/mm/fault.c:479
    do_dat_exception+0x34/0x80 arch/s390/mm/fault.c:567
    pgm_check_handler+0x1da/0x22c arch/s390/kernel/entry.S:706
    copy_from_user_mvcos arch/s390/lib/uaccess.c:111 [inline]
    raw_copy_from_user+0x3a/0x88 arch/s390/lib/uaccess.c:174
    _copy_from_user+0x48/0xa8 lib/usercopy.c:16
    copy_from_user include/linux/uaccess.h:192 [inline]
    __do_sys_sigaltstack kernel/signal.c:4064 [inline]
    __s390x_sys_sigaltstack+0xc8/0x240 kernel/signal.c:4060
    system_call+0xe0/0x28c arch/s390/kernel/entry.S:415

  Allocated by task 9334:
    slab_alloc_node mm/slub.c:2891 [inline]
    slab_alloc mm/slub.c:2899 [inline]
    kmem_cache_alloc+0x118/0x348 mm/slub.c:2904
    vm_area_dup+0x9c/0x2b8 kernel/fork.c:356
    __split_vma+0xba/0x560 mm/mmap.c:2742
    split_vma+0xca/0x108 mm/mmap.c:2800
    mlock_fixup+0x4ae/0x600 mm/mlock.c:550
    apply_vma_lock_flags+0x2c6/0x398 mm/mlock.c:619
    do_mlock+0x1aa/0x718 mm/mlock.c:711
    __do_sys_mlock2 mm/mlock.c:738 [inline]
    __s390x_sys_mlock2+0x86/0xa8 mm/mlock.c:728
    system_call+0xe0/0x28c arch/s390/kernel/entry.S:415

  Freed by task 9333:
    slab_free mm/slub.c:3142 [inline]
    kmem_cache_free+0x7c/0x4b8 mm/slub.c:3158
    __vma_adjust+0x7b2/0x2508 mm/mmap.c:960
    vma_merge+0x87e/0xce0 mm/mmap.c:1209
    userfaultfd_release+0x412/0x6b8 fs/userfaultfd.c:868
    __fput+0x22c/0x7a8 fs/file_table.c:281
    task_work_run+0x200/0x320 kernel/task_work.c:151
    tracehook_notify_resume include/linux/tracehook.h:188 [inline]
    do_notify_resume+0x100/0x148 arch/s390/kernel/signal.c:538
    system_call+0xe6/0x28c arch/s390/kernel/entry.S:416

  The buggy address belongs to the object at 00000000962d6948 which belongs to the cache vm_area_struct of size 200
  The buggy address is located 64 bytes inside of 200-byte region [00000000962d694800000000962d6a10)
  The buggy address belongs to the page: page:00000000313a09fe refcount:1 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x962d6 flags: 0x3ffff00000000200(slab)
  raw: 3ffff00000000200 000040000257e080 0000000c0000000c 000000008020ba00
  raw: 0000000000000000 000f001e00000000 ffffffff00000001 0000000096959501
  page dumped because: kasan: bad access detected
  page->mem_cgroup:0000000096959501

  Memory state around the buggy address:
   00000000962d6880: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
   00000000962d6900: 00 fc fc fc fc fc fc fc fc fa fb fb fb fb fb fb
  >00000000962d6980: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
                        ^
   00000000962d6a00: fb fb fc fc fc fc fc fc fc fc 00 00 00 00 00 00
   00000000962d6a80: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
  ==================================================================

Fixes: 6b251fc96cf2c ("userfaultfd: call handle_userfault() for userfaultfd_missing() faults")
Reported-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Gerald Schaefer <gerald.schaefer@linux.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Heiko Carstens <hca@linux.ibm.com>
Cc: <stable@vger.kernel.org> [4.3+]
Link: https://lkml.kernel.org/r/20201110190329.11920-1-gerald.schaefer@linux.ibm.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
3 years agomm: memcg/slab: fix root memcg vmstats
Muchun Song [Sun, 22 Nov 2020 06:17:12 +0000 (22:17 -0800)]
mm: memcg/slab: fix root memcg vmstats

If we reparent the slab objects to the root memcg, when we free the slab
object, we need to update the per-memcg vmstats to keep it correct for
the root memcg.  Now this at least affects the vmstat of
NR_KERNEL_STACK_KB for !CONFIG_VMAP_STACK when the thread stack size is
smaller than the PAGE_SIZE.

David said:
 "I assume that without this fix that the root memcg's vmstat would
  always be inflated if we reparented"

Fixes: ec9f02384f60 ("mm: workingset: fix vmstat counters for shadow nodes")
Signed-off-by: Muchun Song <songmuchun@bytedance.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Shakeel Butt <shakeelb@google.com>
Acked-by: Roman Gushchin <guro@fb.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: David Rientjes <rientjes@google.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Vladimir Davydov <vdavydov.dev@gmail.com>
Cc: Christopher Lameter <cl@linux.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Roman Gushchin <guro@fb.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Yafang Shao <laoar.shao@gmail.com>
Cc: Chris Down <chris@chrisdown.name>
Cc: <stable@vger.kernel.org> [5.3+]
Link: https://lkml.kernel.org/r/20201110031015.15715-1-songmuchun@bytedance.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
3 years agomm: fix readahead_page_batch for retry entries
Matthew Wilcox (Oracle) [Sun, 22 Nov 2020 06:17:08 +0000 (22:17 -0800)]
mm: fix readahead_page_batch for retry entries

Both btrfs and fuse have reported faults caused by seeing a retry entry
instead of the page they were looking for.  This was caused by a missing
check in the iterator.

As can be seen in the below panic log, the accessing 0x402 causes a
panic.  In the xarray.h, 0x402 means RETRY_ENTRY.

  BUG: kernel NULL pointer dereference, address: 0000000000000402
  CPU: 14 PID: 306003 Comm: as Not tainted 5.9.0-1-amd64 #1 Debian 5.9.1-1
  Hardware name: Lenovo ThinkSystem SR665/7D2VCTO1WW, BIOS D8E106Q-1.01 05/30/2020
  RIP: 0010:fuse_readahead+0x152/0x470 [fuse]
  Code: 41 8b 57 18 4c 8d 54 10 ff 4c 89 d6 48 8d 7c 24 10 e8 d2 e3 28 f9 48 85 c0 0f 84 fe 00 00 00 44 89 f2 49 89 04 d4 44 8d 72 01 <48> 8b 10 41 8b 4f 1c 48 c1 ea 10 83 e2 01 80 fa 01 19 d2 81 e2 01
  RSP: 0018:ffffad99ceaebc50 EFLAGS: 00010246
  RAX: 0000000000000402 RBX: 0000000000000001 RCX: 0000000000000002
  RDX: 0000000000000000 RSI: ffff94c5af90bd98 RDI: ffffad99ceaebc60
  RBP: ffff94ddc1749a00 R08: 0000000000000402 R09: 0000000000000000
  R10: 0000000000000000 R11: 0000000000000100 R12: ffff94de6c429ce0
  R13: ffff94de6c4d3700 R14: 0000000000000001 R15: ffffad99ceaebd68
  FS:  00007f228c5c7040(0000) GS:ffff94de8ed80000(0000) knlGS:0000000000000000
  CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
  CR2: 0000000000000402 CR3: 0000001dbd9b4000 CR4: 0000000000350ee0
  Call Trace:
    read_pages+0x83/0x270
    page_cache_readahead_unbounded+0x197/0x230
    generic_file_buffered_read+0x57a/0xa20
    new_sync_read+0x112/0x1a0
    vfs_read+0xf8/0x180
    ksys_read+0x5f/0xe0
    do_syscall_64+0x33/0x80
    entry_SYSCALL_64_after_hwframe+0x44/0xa9

Fixes: 042124cc64c3 ("mm: add new readahead_control API")
Reported-by: David Sterba <dsterba@suse.com>
Reported-by: Wonhyuk Yang <vvghjk1234@gmail.com>
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: <stable@vger.kernel.org>
Link: https://lkml.kernel.org/r/20201103142852.8543-1-willy@infradead.org
Link: https://lkml.kernel.org/r/20201103124349.16722-1-vvghjk1234@gmail.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
3 years agomm: fix phys_to_target_node() and memory_add_physaddr_to_nid() exports
Dan Williams [Sun, 22 Nov 2020 06:17:05 +0000 (22:17 -0800)]
mm: fix phys_to_target_node() and memory_add_physaddr_to_nid() exports

The core-mm has a default __weak implementation of phys_to_target_node()
to mirror the weak definition of memory_add_physaddr_to_nid().  That
symbol is exported for modules.  However, while the export in
mm/memory_hotplug.c exported the symbol in the configuration cases of:

CONFIG_NUMA_KEEP_MEMINFO=y
CONFIG_MEMORY_HOTPLUG=y

...and:

CONFIG_NUMA_KEEP_MEMINFO=n
CONFIG_MEMORY_HOTPLUG=y

...it failed to export the symbol in the case of:

CONFIG_NUMA_KEEP_MEMINFO=y
CONFIG_MEMORY_HOTPLUG=n

Not only is that broken, but Christoph points out that the kernel should
not be exporting any __weak symbol, which means that
memory_add_physaddr_to_nid() example that phys_to_target_node() copied
is broken too.

Rework the definition of phys_to_target_node() and
memory_add_physaddr_to_nid() to not require weak symbols.  Move to the
common arch override design-pattern of an asm header defining a symbol
to replace the default implementation.

The only common header that all memory_add_physaddr_to_nid() producing
architectures implement is asm/sparsemem.h.  In fact, powerpc already
defines its memory_add_physaddr_to_nid() helper in sparsemem.h.
Double-down on that observation and define phys_to_target_node() where
necessary in asm/sparsemem.h.  An alternate consideration that was
discarded was to put this override in asm/numa.h, but that entangles
with the definition of MAX_NUMNODES relative to the inclusion of
linux/nodemask.h, and requires powerpc to grow a new header.

The dependency on NUMA_KEEP_MEMINFO for DEV_DAX_HMEM_DEVICES is invalid
now that the symbol is properly exported / stubbed in all combinations
of CONFIG_NUMA_KEEP_MEMINFO and CONFIG_MEMORY_HOTPLUG.

[dan.j.williams@intel.com: v4]
Link: https://lkml.kernel.org/r/160461461867.1505359.5301571728749534585.stgit@dwillia2-desk3.amr.corp.intel.com
[dan.j.williams@intel.com: powerpc: fix create_section_mapping compile warning]
Link: https://lkml.kernel.org/r/160558386174.2948926.2740149041249041764.stgit@dwillia2-desk3.amr.corp.intel.com
Fixes: a035b6bf863e ("mm/memory_hotplug: introduce default phys_to_target_node() implementation")
Reported-by: Randy Dunlap <rdunlap@infradead.org>
Reported-by: Thomas Gleixner <tglx@linutronix.de>
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Tested-by: Randy Dunlap <rdunlap@infradead.org>
Tested-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Cc: Joao Martins <joao.m.martins@oracle.com>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Vishal Verma <vishal.l.verma@intel.com>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Link: https://lkml.kernel.org/r/160447639846.1133764.7044090803980177548.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>