linux-stable.git/sound, branch v4.20.3

ALSA: hda/realtek - Disable headset Mic VREF for headset mode of ALC225

2019-01-16T21:03:21+00:00

commit d1dd42110d2727e81b9265841a62bc84c454c3a2 upstream.

Disable Headset Mic VREF for headset mode of ALC225.
This will be controlled by coef bits of headset mode functions.

[ Fixed a compile warning and code simplification -- tiwai ]

Signed-off-by: Kailang Yang 
Cc: 
Signed-off-by: Takashi Iwai 
Signed-off-by: Greg Kroah-Hartman

ALSA: hda/realtek - Add unplug function into unplug state of Headset Mode for ALC225

2019-01-16T21:03:21+00:00

commit 4d4b0c52bde470c379f5d168d5c139ad866cb808 upstream.

Forgot to add unplug function to unplug state of headset mode
for ALC225.

Signed-off-by: Kailang Yang 
Cc: 
Signed-off-by: Takashi Iwai 
Signed-off-by: Greg Kroah-Hartman

ALSA: hda/realtek - Support Dell headset mode for New AIO platform

2019-01-16T21:03:21+00:00

commit c2a7c55a04065c3b0c32d23b099db7ea1dbf6250 upstream.

Dell has new platform for ALC274.
This will support to enable headset mode.

Signed-off-by: Kailang Yang 
Cc: 
Signed-off-by: Takashi Iwai 
Signed-off-by: Greg Kroah-Hartman

ALSA: usb-audio: Always check descriptor sizes in parser code

2019-01-13T08:24:05+00:00

commit 3e96d7280f16e2f787307f695a31296b9e4a1cd7 upstream.

There are a few places where we access the data without checking the
actual object size from the USB audio descriptor.  This may result in
OOB access, as recently reported.

This patch addresses these missing checks.  Most of added codes are
simple bLength checks in the caller side.  For the input and output
terminal parsers, we put the length check in the parser functions.
For the input terminal, a new argument is added to distinguish between
UAC1 and the rest, as they treat different objects.

Reported-by: Mathias Payer 
Reported-by: Hui Peng 
Tested-by: Hui Peng 
Cc: 
Signed-off-by: Takashi Iwai 
Signed-off-by: Greg Kroah-Hartman

ALSA: usb-audio: Fix an out-of-bound read in create_composite_quirks

2019-01-13T08:24:04+00:00

commit cbb2ebf70daf7f7d97d3811a2ff8e39655b8c184 upstream.

In `create_composite_quirk`, the terminating condition of for loops is
`quirk->ifnum < 0`. So any composite quirks should end with `struct
snd_usb_audio_quirk` object with ifnum < 0.

    for (quirk = quirk_comp->data; quirk->ifnum >= 0; ++quirk) {

    	.....
    }

the data field of Bower's & Wilkins PX headphones usb device device quirks
do not end with {.ifnum = -1}, wihch may result in out-of-bound read.

This Patch fix the bug by adding an ending quirk object.

Fixes: 240a8af929c7 ("ALSA: usb-audio: Add a quirck for B&W PX headphones")
Signed-off-by: Hui Peng 
Cc: 
Signed-off-by: Takashi Iwai 
Signed-off-by: Greg Kroah-Hartman

ALSA: usb-audio: Check mixer unit descriptors more strictly

2019-01-13T08:24:04+00:00

commit 0bfe5e434e6665b3590575ec3c5e4f86a1ce51c9 upstream.

We've had some sanity checks of the mixer unit descriptors but they
are too loose and some corner cases are overlooked.  Add more strict
checks in uac_mixer_unit_get_channels() for avoiding possible OOB
accesses by malformed descriptors.

This also changes the semantics of uac_mixer_unit_get_channels()
slightly.  Now it returns zero for the cases where the descriptor
lacks of bmControls instead of -EINVAL.  Then the caller side skips
the mixer creation for such unit while it keeps parsing it.
This corresponds to the case like Maya44.

Cc: 
Signed-off-by: Takashi Iwai 
Signed-off-by: Greg Kroah-Hartman

ALSA: usb-audio: Avoid access before bLength check in build_audio_procunit()

2019-01-13T08:24:04+00:00

commit f4351a199cc120ff9d59e06d02e8657d08e6cc46 upstream.

The parser for the processing unit reads bNrInPins field before the
bLength sanity check, which may lead to an out-of-bound access when a
malformed descriptor is given.  Fix it by assignment after the bLength
check.

Cc: 
Signed-off-by: Takashi Iwai 
Signed-off-by: Greg Kroah-Hartman

ALSA: cs46xx: Potential NULL dereference in probe

2019-01-13T08:24:04+00:00

commit 1524f4e47f90b27a3ac84efbdd94c63172246a6f upstream.

The "chip->dsp_spos_instance" can be NULL on some of the ealier error
paths in snd_cs46xx_create().

Reported-by: "Yavuz, Tuba" 
Signed-off-by: Dan Carpenter 
Cc: 
Signed-off-by: Takashi Iwai 
Signed-off-by: Greg Kroah-Hartman

ASoC: intel: cht_bsw_max98090_ti: Add pmc_plt_clk_0 quirk for Chromebook Gnawty

2019-01-09T16:45:54+00:00

commit 94ea56cff506c769a509c5dd87904c7fe3806a81 upstream.

The Gnawty model Chromebook uses pmc_plt_clk_0 instead of pmc_plt_clk_3
for the mclk, just like the Clapper and Swanky models.

This commit adds a DMI based quirk for this.

This fixing audio no longer working on these devices after
commit 648e921888ad ("clk: x86: Stop marking clocks as CLK_IS_CRITICAL")
that commit fixes us unnecessary keeping unused clocks on, but in case of
the Gnawty that was breaking audio support since we were not using the
right clock in the cht_bsw_max98090_ti machine driver.

BugLink: https://bugzilla.kernel.org/show_bug.cgi?id=201787
Cc: stable@vger.kernel.org
Fixes: 648e921888ad ("clk: x86: Stop marking clocks as CLK_IS_CRITICAL")
Reported-and-tested-by: Jaime Pérez <19.jaime.91@gmail.com>
Signed-off-by: Hans de Goede 
Acked-by: Pierre-Louis Bossart 
Signed-off-by: Mark Brown 
Signed-off-by: Greg Kroah-Hartman

ASoC: intel: cht_bsw_max98090_ti: Add pmc_plt_clk_0 quirk for Chromebook Clapper

2019-01-09T16:45:54+00:00

commit 984bfb398a3af6fa9b7e80165e524933b0616686 upstream.

The Clapper model Chromebook uses pmc_plt_clk_0 instead of pmc_plt_clk_3
for the mclk, just like the Swanky model.

This commit adds a DMI based quirk for this.

This fixing audio no longer working on these devices after
commit 648e921888ad ("clk: x86: Stop marking clocks as CLK_IS_CRITICAL")
that commit fixes us unnecessary keeping unused clocks on, but in case of
the Clapper that was breaking audio support since we were not using the
right clock in the cht_bsw_max98090_ti machine driver.

Cc: stable@vger.kernel.org
Fixes: 648e921888ad ("clk: x86: Stop marking clocks as CLK_IS_CRITICAL")
Signed-off-by: Hans de Goede 
Acked-by: Pierre-Louis Bossart 
Signed-off-by: Mark Brown 
Signed-off-by: Greg Kroah-Hartman