platform/kernel/linux-rpi.git
5 years agoMerge branch 'asoc-5.4' into asoc-next
Mark Brown [Sun, 15 Sep 2019 09:31:44 +0000 (10:31 +0100)]
Merge branch 'asoc-5.4' into asoc-next

5 years agoMerge branch 'asoc-5.3' into asoc-linus
Mark Brown [Sun, 15 Sep 2019 09:31:42 +0000 (10:31 +0100)]
Merge branch 'asoc-5.3' into asoc-linus

5 years agoASoC: sdm845: remove unneeded semicolon
Saiyam Doshi [Sat, 14 Sep 2019 03:11:33 +0000 (08:41 +0530)]
ASoC: sdm845: remove unneeded semicolon

Remove excess semicolon after closing parenthesis.

Signed-off-by: Saiyam Doshi <saiyamdoshi.in@gmail.com>
Link: https://lore.kernel.org/r/20190914031133.GA28447@SD
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: fsl_sai: Implement set_bclk_ratio
Viorel Suman [Fri, 30 Aug 2019 21:59:10 +0000 (00:59 +0300)]
ASoC: fsl_sai: Implement set_bclk_ratio

This is to allow machine drivers to set a certain bitclk rate
which might not be exactly rate * frame size.

Cc: NXP Linux Team <linux-imx@nxp.com>
Signed-off-by: Viorel Suman <viorel.suman@nxp.com>
Signed-off-by: Daniel Baluta <daniel.baluta@nxp.com>
Acked-by: Nicolin Chen <nicoleotsuka@gmail.com>
Link: https://lore.kernel.org/r/20190830215910.31590-1-daniel.baluta@nxp.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: dmaengine: Replace strncpy() with strscpy_pad() for pcm->name
Peter Ujfalusi [Wed, 11 Sep 2019 08:33:31 +0000 (11:33 +0300)]
ASoC: dmaengine: Replace strncpy() with strscpy_pad() for pcm->name

While it is safe to use strncpy in this case, the advice is to move to
strscpy or strscpy_pad.

Suggested-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Link: https://lore.kernel.org/r/20190911083331.16801-1-peter.ujfalusi@ti.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: wcd9335: remove redundant use of ret variable
Saiyam Doshi [Mon, 9 Sep 2019 17:45:41 +0000 (23:15 +0530)]
ASoC: wcd9335: remove redundant use of ret variable

All these functions declares and initializes variable ret with
'0' and without modifying 'ret' variable, it is returned.

This patch removes this redundancy and returns '0' directly.

Signed-off-by: Saiyam Doshi <saiyamdoshi.in@gmail.com>
Link: https://lore.kernel.org/r/20190909174541.GA22718@SD
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: hdmi-codec: Add an op to set callback function for plug event
Cheng-Yi Chiang [Wed, 17 Jul 2019 08:33:23 +0000 (16:33 +0800)]
ASoC: hdmi-codec: Add an op to set callback function for plug event

Add an op in hdmi_codec_ops so codec driver can register callback
function to handle plug event.

Driver in DRM can use this callback function to report connector status.

Signed-off-by: Cheng-Yi Chiang <cychiang@chromium.org>
Link: https://lore.kernel.org/r/20190717083327.47646-2-cychiang@chromium.org
Reviewed-by: Tzung-Bi Shih <tzungbi@google.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: rt5677: keep analog power register at SND_SOC_BIAS_OFF
Ben Zhang [Fri, 6 Sep 2019 19:46:24 +0000 (12:46 -0700)]
ASoC: rt5677: keep analog power register at SND_SOC_BIAS_OFF

Instead of clearing RT5677_PWR_ANLG2 (MX-64h) to 0 at SND_SOC_BIAS_OFF,
we only clear the RT5677_PWR_CORE bit which is set at SND_SOC_BIAS_PREPARE.
MICBIAS control bits are left unchanged.

This fixed the bug where if MICBIAS1 widget is forced on, MICBIAS
control bits will be cleared at suspend and never turned back on again,
since DAPM thinks the widget is always on.

Signed-off-by: Ben Zhang <benzh@chromium.org>
Signed-off-by: Curtis Malainey <cujomalainey@chromium.org>
Link: https://lore.kernel.org/r/20190906194636.217881-3-cujomalainey@chromium.org
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: rt5677: Remove magic number register writes
Curtis Malainey [Fri, 6 Sep 2019 19:46:23 +0000 (12:46 -0700)]
ASoC: rt5677: Remove magic number register writes

In order to simplify understanding what register values are being
written to the codec for debugging more advanced features (such as
hotwording) it is best to remove magic numbers

Signed-off-by: Curtis Malainey <cujomalainey@chromium.org>
Link: https://lore.kernel.org/r/20190906194636.217881-2-cujomalainey@chromium.org
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: soc-core: self contained soc_unbind_aux_dev()
Kuninori Morimoto [Wed, 4 Sep 2019 00:15:40 +0000 (09:15 +0900)]
ASoC: soc-core: self contained soc_unbind_aux_dev()

Current soc_unbind_aux_dev() implementation is very half,
thus it is very unreadable.

for_each_comp_order(order) {
for_each_card_auxs_safe(card, comp, _comp) {

(1) if (comp->driver->remove_order == order) {
...
=> soc_unbind_aux_dev(comp);
}
}

soc_unbind_aux_dev() itself is not related to remove_order (1).
And, it is called from soc_remove_aux_devices(), even though
its paired function soc_bind_aux_dev() is called from
snd_soc_instantiate_card().
It is very unbalance, and very difficult to understand.

This patch do
1) update soc_bind_aux_dev() to self contained
2) call it from soc_cleanup_card_resources() to make up balance

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/87r24wor0z.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: soc-core: add soc_unbind_aux_dev()
Kuninori Morimoto [Wed, 4 Sep 2019 00:15:35 +0000 (09:15 +0900)]
ASoC: soc-core: add soc_unbind_aux_dev()

It is easy to read code if it is cleanly using paired function/naming,
like start <-> stop, register <-> unregister, etc, etc.
But, current ALSA SoC code is very random, unbalance, not paired, etc.
It is easy to create bug at the such code, and it will be difficult to
debug.

soc-core.c has soc_bind_aux_dev(), but, there is no its paired
soc_unbind_aux_dev().
This patch adds soc_unbind_aux_dev().

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/87sgpcor14.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: soc-core: self contained soc_bind_aux_dev()
Kuninori Morimoto [Wed, 4 Sep 2019 00:15:28 +0000 (09:15 +0900)]
ASoC: soc-core: self contained soc_bind_aux_dev()

Current soc_bind_aux_dev() implementation is very half,
thus it is very unreadable.

for_each_card_pre_auxs(xxx) {
=> ret = soc_bind_aux_dev(xxx);
...
}

This patch does all for_each_xxx() under soc_bind_aux_dev(),
and makes it to self contained.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/87tv9sor1b.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: soc-core: move soc_probe_link_dais() next to soc_remove_link_dais()
Kuninori Morimoto [Wed, 4 Sep 2019 00:15:23 +0000 (09:15 +0900)]
ASoC: soc-core: move soc_probe_link_dais() next to soc_remove_link_dais()

It is easy to read code if it is cleanly using paired function/naming,
like start <-> stop, register <-> unregister, etc, etc.
But, current ALSA SoC code is very random, unbalance, not paired, etc.
It is easy to create bug at the such code, and it will be difficult to
debug.

This patch moves soc_probe_link_dais() next to soc_remove_link_dais()
which is paired function.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/87v9u8or1g.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: soc-core: self contained soc_probe_link_dais()
Kuninori Morimoto [Wed, 4 Sep 2019 00:15:17 +0000 (09:15 +0900)]
ASoC: soc-core: self contained soc_probe_link_dais()

Current soc_probe_link_dais() implementation is very half,
thus it is very difficult to read.

for_each_comp_order(xxx) {
for_each_card_rtds(xxx)
=> soc_probe_link_dais(xxx);
}

This patch does all for_each_xxx() under soc_probe_link_dais(),
and makes it to self contained.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/87woeoor1m.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: soc-core: add new soc_link_init()
Kuninori Morimoto [Wed, 4 Sep 2019 00:15:12 +0000 (09:15 +0900)]
ASoC: soc-core: add new soc_link_init()

Current soc_probe_link_dais() (1) is called under probe_order (2),
and it will initialize dai_link related settings at *Last* turn (3)(B).
It is very complex code.

static int soc_probe_link_dais(..., order)
{
(A) /* probe DAIs here */
...

(3) if (order != SND_SOC_COMP_ORDER_LAST)
return 0;

(B) /* initialize dai_link related settings */
...
}

static int snd_soc_instantiate_card(...)
{
...
(2) for_each_comp_order(order) {
for_each_card_rtds(...) {
(1) ret = soc_probe_link_dais(..., order);
}
}
}

This patch separes soc_probe_link_dais() into "DAI probe" portion (A),
and dai_link settings portion (B).
The later is named as soc_link_init() by this patch.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/87y2z4or1r.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: soc-core: move soc_probe_dai() next to soc_remove_dai()
Kuninori Morimoto [Wed, 4 Sep 2019 00:15:06 +0000 (09:15 +0900)]
ASoC: soc-core: move soc_probe_dai() next to soc_remove_dai()

It is easy to read code if it is cleanly using paired function/naming,
like start <-> stop, register <-> unregister, etc, etc.
But, current ALSA SoC code is very random, unbalance, not paired, etc.
It is easy to create bug at the such code, and it will be difficult to
debug.

This patch moves soc_probe_dai() next to soc_remove_dai() which is
paired function.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/87zhjkor1x.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: soc-core: self contained soc_remove_link_dais()
Kuninori Morimoto [Wed, 4 Sep 2019 00:14:57 +0000 (09:14 +0900)]
ASoC: soc-core: self contained soc_remove_link_dais()

Current soc_remove_link_dais() implementation is very half,
thus it is very difficult to read.

for_each_comp_order(xxx) {
for_each_card_rtds(xxx)
=> soc_remove_link_dais(xxx);
}

This patch does all for_each_xxx() under soc_remove_link_dais(),
and makes it to self contained.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/871rwwq5mm.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: soc-core: self contained soc_remove_link_components()
Kuninori Morimoto [Wed, 4 Sep 2019 00:14:51 +0000 (09:14 +0900)]
ASoC: soc-core: self contained soc_remove_link_components()

Current soc_remove_link_components() implementation is very half,
thus it is very difficult to read.

for_each_comp_order(xxx) {
for_each_card_rtds(xxx)
=> soc_remove_link_components(xxx);
}

This patch does all for_each_xxx() under soc_remove_link_components(),
and makes it to self contained.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/8736hcq5ms.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: soc-core: self contained soc_probe_link_components()
Kuninori Morimoto [Wed, 4 Sep 2019 00:14:46 +0000 (09:14 +0900)]
ASoC: soc-core: self contained soc_probe_link_components()

Current soc_probe_link_components() implementation is very half,
thus it is very difficult to read.

for_each_comp_order(xxx) {
for_each_card_rtds(xxx) {
=> ret = soc_probe_link_components(xxx);
...
}
}

This patch does all for_each_xxx() under soc_probe_link_components(),
and makes it to self contained.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/874l1sq5mx.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: rt1308: make array pd static const, makes object smaller
Colin Ian King [Sat, 7 Sep 2019 07:46:34 +0000 (08:46 +0100)]
ASoC: rt1308: make array pd static const, makes object smaller

Don't populate the array pd on the stack but instead make it
static const. Makes the object code smaller by 82 bytes.

Before:
   text    data     bss     dec     hex filename
  26548    7288      64   33900    846c sound/soc/codecs/rt1308.o

After:
   text    data     bss     dec     hex filename
  26370    7384      64   33818    841a sound/soc/codecs/rt1308.o

(gcc version 9.2.1, amd64)

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Link: https://lore.kernel.org/r/20190907074634.22144-1-colin.king@canonical.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: rt1305: make array pd static const, makes object smaller
Colin Ian King [Sat, 7 Sep 2019 07:41:56 +0000 (08:41 +0100)]
ASoC: rt1305: make array pd static const, makes object smaller

Don't populate the array pd on the stack but instead make it
static const. Makes the object code smaller by 93 bytes.

Before:
   text    data     bss     dec     hex filename
  38961    9784      64   48809    bea9 sound/soc/codecs/rt1305.o

After:
   text    data     bss     dec     hex filename
  38804    9848      64   48716    be4c sound/soc/codecs/rt1305.o

(gcc version 9.2.1, amd64)

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Link: https://lore.kernel.org/r/20190907074156.21907-1-colin.king@canonical.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: rt1011: make array pd static const, makes object smaller
Colin Ian King [Sat, 7 Sep 2019 07:37:17 +0000 (08:37 +0100)]
ASoC: rt1011: make array pd static const, makes object smaller

Don't populate the array pd on the stack but instead make it
static const. Makes the object code smaller by 100 bytes.

Before:
   text    data     bss     dec     hex filename
  51463   13016     128   64607    fc5f sound/soc/codecs/rt1011.o

After:
   text    data     bss     dec     hex filename
  51299   13080     128   64507    fbfb sound/soc/codecs/rt1011.o

(gcc version 9.2.1, amd64)

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Link: https://lore.kernel.org/r/20190907073717.21632-1-colin.king@canonical.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoSoC: simple-card-utils: set 0Hz to sysclk when shutdown
Katsuhiro Suzuki [Sat, 7 Sep 2019 17:45:01 +0000 (02:45 +0900)]
SoC: simple-card-utils: set 0Hz to sysclk when shutdown

This patch set 0Hz to sysclk when shutdown the card.

Some codecs set rate constraints that derives from sysclk. This
mechanism works correctly if machine drivers give fixed frequency.

But simple-audio and audio-graph card set variable clock rate if
'mclk-fs' property exists. In this case, rate constraints will go
bad scenario. For example a codec accepts three limited rates
(mclk / 256, mclk / 384, mclk / 512).

Bad scenario as follows (mclk-fs = 256):
   - Initialize sysclk by correct value (Ex. 12.288MHz)
     - Codec set constraints of PCM rate by sysclk
       48kHz (1/256), 32kHz (1/384), 24kHz (1/512)
   - Play 48kHz sound, it's acceptable
   - Sysclk is not changed

   - Play 32kHz sound, it's acceptable
   - Set sysclk to 8.192MHz (= fs * mclk-fs = 32k * 256)
     - Codec set constraints of PCM rate by sysclk
       32kHz (1/256), 21.33kHz (1/384), 16kHz (1/512)

   - Play 48kHz again, but it's NOT acceptable because constraints
     do not allow 48kHz

So codecs treat 0Hz sysclk as signal of applying no constraints to
avoid this problem.

Signed-off-by: Katsuhiro Suzuki <katsuhiro@katsuster.net>
Link: https://lore.kernel.org/r/20190907174501.19833-1-katsuhiro@katsuster.net
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: rockchip: ignore 0Hz sysclk
Katsuhiro Suzuki [Sat, 7 Sep 2019 17:43:32 +0000 (02:43 +0900)]
ASoC: rockchip: ignore 0Hz sysclk

This patch ignores sysclk setting if it is 0Hz.

Some codecs treat 0Hz sysclk as signal of applying no constraints.
This driver does not have such feature but current implementation
outputs 'Failed to set mclk' error message if machine driver sets
0Hz sysclk to this driver.

Signed-off-by: Katsuhiro Suzuki <katsuhiro@katsuster.net>
Link: https://lore.kernel.org/r/20190907174332.19586-1-katsuhiro@katsuster.net
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: codecs: ad193x: make two arrays static const, makes object smaller
Colin Ian King [Fri, 6 Sep 2019 16:14:04 +0000 (17:14 +0100)]
ASoC: codecs: ad193x: make two arrays static const, makes object smaller

Don't populate the arrays on the stack but instead make them
static const. Makes the object code smaller by 37 bytes.

Before:
   text    data     bss     dec     hex filename
  16253    7200       0   23453    5b9d sound/soc/codecs/ad193x.o

After:
   text    data     bss     dec     hex filename
  16056    7360       0   23416    5b78 sound/soc/codecs/ad193x.o

(gcc version 9.2.1, amd64)

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Link: https://lore.kernel.org/r/20190906161404.1440-1-colin.king@canonical.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: es8316: support fixed and variable both clock rates
Katsuhiro Suzuki [Sat, 7 Sep 2019 16:36:53 +0000 (01:36 +0900)]
ASoC: es8316: support fixed and variable both clock rates

This patch supports some type of machine drivers that set 0 to mclk
when sound device goes to idle state. After applied this patch,
sysclk == 0 means there is no constraint of sound rate and other
values will set constraints which is derived by sysclk setting.

Original code refuses sysclk == 0 setting. But some boards and SoC
(such as RockPro64 and RockChip I2S) has connected SoC MCLK out to
ES8316 MCLK in. In this case, SoC side I2S will choose suitable
frequency of MCLK such as fs * mclk-fs when user starts playing or
capturing.

Bad scenario as follows (mclk-fs = 256):
  - Initialize sysclk by correct value (Ex. 12.288MHz)
    - ES8316 set constraints of PCM rate by sysclk
      48kHz (1/256), 32kHz (1/384), 30.720kHz (1/400),
      24kHz (1/512), 16kHz (1/768), 12kHz (1/1024)
  - Play 48kHz sound, it's acceptable
  - Sysclk is not changed

  - Play 32kHz sound, it's acceptable
  - Set sysclk by 8.192MHz (= fs * mclk-fs = 32k * 256)
    - ES8316 set constraints of PCM rate by sysclk
      32kHz (1/256), 21.33kHz (1/384), 20.48kHz (1/400),
      16kHz (1/512), 10.66kHz (1/768), 8kHz (1/1024)

  - Play 48kHz again, but it's NOT acceptable because constraints
    list does not allow 48kHz

Signed-off-by: Katsuhiro Suzuki <katsuhiro@katsuster.net>
Link: https://lore.kernel.org/r/20190907163653.9382-2-katsuhiro@katsuster.net
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: es8316: fix redundant codes of clock
Katsuhiro Suzuki [Sat, 7 Sep 2019 16:36:52 +0000 (01:36 +0900)]
ASoC: es8316: fix redundant codes of clock

This patch removes redundant null checks for optional MCLK clock.
And fix DT binding document for changing clock property to optional
from required.

Signed-off-by: Katsuhiro Suzuki <katsuhiro@katsuster.net>
Link: https://lore.kernel.org/r/20190907163653.9382-1-katsuhiro@katsuster.net
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: ams-delta: Take control over audio mute GPIO pins
Janusz Krzysztofik [Sat, 7 Sep 2019 11:16:50 +0000 (13:16 +0200)]
ASoC: ams-delta: Take control over audio mute GPIO pins

Since commit 1137ceee76ba ("ARM: OMAP1: ams-delta: Don't request unused
GPIOs"), on-board audio has appeared muted.  It has been discovered that
believed to be unused GPIO pins "hookflash1" and "hookflash2" need to be
set low for audible sound in handsfree and handset mode respectively.

According to Amstrad E3 wiki, the purpose of both pins hasn't been
clearly identified.  Original Amstrad software used to produce a high
pulse on them when the phone was taken off hook or recall was pressed.
With the current findings, we can assume the pins provide a kind of
audio mute function, separately for handset and handsfree operation
modes.

Commit 2afdb4c41d78 ("ARM: OMAP1: ams-delta: Fix audio permanently
muted") attempted to fix the issue temporarily by hogging the GPIO pin
"hookflash1" renamed to "audio_mute", however the fix occurred
incomplete as it restored audible sound only for handsfree mode.

Stop hogging that pin, rename the pins to "handsfree_mute" and
"handset_mute" respectively and implement appropriate DAPM event
callbacks for "Speaker" and "Earpiece" DAPM widgets.

Fixes: 1137ceee76ba ("ARM: OMAP1: ams-delta: Don't request unused GPIOs")
Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com>
Reviewed-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Link: https://lore.kernel.org/r/20190907111650.15440-1-jmkrzyszt@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: dmaengine: Make the pcm->name equal to pcm->id if the name is not set
Peter Ujfalusi [Fri, 6 Sep 2019 05:55:24 +0000 (08:55 +0300)]
ASoC: dmaengine: Make the pcm->name equal to pcm->id if the name is not set

Some tools use the snd_pcm_info_get_name() to try to identify PCMs or for
other purposes.

Currently it is left empty with the dmaengine-pcm, in this case copy the
pcm->id string as pcm->name.

For example IGT is using this to find the HDMI PCM for testing audio on it.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Reported-by: Arthur She <arthur.she@linaro.org>
Link: https://lore.kernel.org/r/20190906055524.7393-1-peter.ujfalusi@ti.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoLinux 5.3-rc8
Linus Torvalds [Sun, 8 Sep 2019 20:33:15 +0000 (13:33 -0700)]
Linux 5.3-rc8

5 years agoMerge tag 'compiler-attributes-for-linus-v5.3-rc8' of git://github.com/ojeda/linux
Linus Torvalds [Sun, 8 Sep 2019 16:34:55 +0000 (09:34 -0700)]
Merge tag 'compiler-attributes-for-linus-v5.3-rc8' of git://github.com/ojeda/linux

Pull section attribute fix from Miguel Ojeda:
 "Fix Oops in Clang-compiled kernels (Nick Desaulniers)"

* tag 'compiler-attributes-for-linus-v5.3-rc8' of git://github.com/ojeda/linux:
  include/linux/compiler.h: fix Oops for Clang-compiled kernels

5 years agoMerge tag 'gpio-v5.3-5' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux...
Linus Torvalds [Sun, 8 Sep 2019 16:30:31 +0000 (09:30 -0700)]
Merge tag 'gpio-v5.3-5' of git://git./linux/kernel/git/linusw/linux-gpio

Pull GPIO fixes from Linus Walleij:
 "All related to the PCA953x driver when handling chips with more than 8
  ports, now that works again"

* tag 'gpio-v5.3-5' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio:
  gpio: pca953x: use pca953x_read_regs instead of regmap_bulk_read
  gpio: pca953x: correct type of reg_direction

5 years agoinclude/linux/compiler.h: fix Oops for Clang-compiled kernels
Nick Desaulniers [Wed, 28 Aug 2019 22:55:23 +0000 (15:55 -0700)]
include/linux/compiler.h: fix Oops for Clang-compiled kernels

GCC unescapes escaped string section names while Clang does not. Because
__section uses the `#` stringification operator for the section name, it
doesn't need to be escaped.

This fixes an Oops observed in distro's that use systemd and not
net.core.bpf_jit_enable=1, when their kernels are compiled with Clang.

Link: https://github.com/ClangBuiltLinux/linux/issues/619
Link: https://bugs.llvm.org/show_bug.cgi?id=42950
Link: https://marc.info/?l=linux-netdev&m=156412960619946&w=2
Link: https://lore.kernel.org/lkml/20190904181740.GA19688@gmail.com/
Acked-by: Will Deacon <will@kernel.org>
Reported-by: Sedat Dilek <sedat.dilek@gmail.com>
Suggested-by: Josh Poimboeuf <jpoimboe@redhat.com>
Tested-by: Sedat Dilek <sedat.dilek@gmail.com>
Signed-off-by: Nick Desaulniers <ndesaulniers@google.com>
[Cherry-picked from the __section cleanup series for 5.3]
[Adjusted commit message]
Signed-off-by: Miguel Ojeda <miguel.ojeda.sandonis@gmail.com>
5 years agoRevert "x86/apic: Include the LDR when clearing out APIC registers"
Linus Torvalds [Sat, 7 Sep 2019 21:25:54 +0000 (14:25 -0700)]
Revert "x86/apic: Include the LDR when clearing out APIC registers"

This reverts commit 558682b5291937a70748d36fd9ba757fb25b99ae.

Chris Wilson reports that it breaks his CPU hotplug test scripts.  In
particular, it breaks offlining and then re-onlining the boot CPU, which
we treat specially (and the BIOS does too).

The symptoms are that we can offline the CPU, but it then does not come
back online again:

    smpboot: CPU 0 is now offline
    smpboot: Booting Node 0 Processor 0 APIC 0x0
    smpboot: do_boot_cpu failed(-1) to wakeup CPU#0

Thomas says he knows why it's broken (my personal suspicion: our magic
handling of the "cpu0_logical_apicid" thing), but for 5.3 the right fix
is to just revert it, since we've never touched the LDR bits before, and
it's not worth the risk to do anything else at this stage.

[ Hotpluging of the boot CPU is special anyway, and should be off by
  default. See the "BOOTPARAM_HOTPLUG_CPU0" config option and the
  cpu0_hotplug kernel parameter.

  In general you should not do it, and it has various known limitations
  (hibernate and suspend require the boot CPU, for example).

  But it should work, even if the boot CPU is special and needs careful
  treatment       - Linus ]

Link: https://lore.kernel.org/lkml/156785100521.13300.14461504732265570003@skylake-alporthouse-com/
Reported-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Bandan Das <bsd@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agoMerge tag 'char-misc-5.3-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Sat, 7 Sep 2019 18:48:28 +0000 (11:48 -0700)]
Merge tag 'char-misc-5.3-rc8' of git://git./linux/kernel/git/gregkh/char-misc

Pull Documentation updates from Greg KH:
 "A few small patches for the documenation file that came in through the
  char-misc tree in -rc7 for your tree.

  They fix the mistake in the .rst format that kept the table of
  companies from showing up in the html output, and most importantly,
  add people's names to the list showing support for our process"

* tag 'char-misc-5.3-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc:
  Documentation/process: Add Qualcomm process ambassador for hardware security issues
  Documentation/process/embargoed-hardware-issues: Microsoft ambassador
  Documentation/process: Add Google contact for embargoed hardware issues
  Documentation/process: Volunteer as the ambassador for Xen

5 years agoDocumentation/process: Add Qualcomm process ambassador for hardware security issues
Trilok Soni [Fri, 6 Sep 2019 19:01:57 +0000 (12:01 -0700)]
Documentation/process: Add Qualcomm process ambassador for hardware security issues

Add Trilok Soni as process ambassador for hardware security issues
from Qualcomm.

Signed-off-by: Trilok Soni <tsoni@codeaurora.org>
Link: https://lore.kernel.org/r/1567796517-8964-1-git-send-email-tsoni@codeaurora.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agoMerge tag 'dmaengine-fix-5.3' of git://git.infradead.org/users/vkoul/slave-dma
Linus Torvalds [Sat, 7 Sep 2019 17:00:34 +0000 (10:00 -0700)]
Merge tag 'dmaengine-fix-5.3' of git://git.infradead.org/users/vkoul/slave-dma

Pull dmaengine fixes from Vinod Koul:
 "Some late fixes for drivers:

   - memory leak in ti crossbar dma driver

   - cleanup of omap dma probe

   - Fix for link list configuration in sprd dma driver

   - Handling fixed for DMACHCLR if iommu is mapped in rcar dma"

* tag 'dmaengine-fix-5.3' of git://git.infradead.org/users/vkoul/slave-dma:
  dmaengine: rcar-dmac: Fix DMACHCLR handling if iommu is mapped
  dmaengine: sprd: Fix the DMA link-list configuration
  dmaengine: ti: omap-dma: Add cleanup in omap_dma_probe()
  dmaengine: ti: dma-crossbar: Fix a memory leak bug

5 years agoMerge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Linus Torvalds [Fri, 6 Sep 2019 23:18:43 +0000 (16:18 -0700)]
Merge tag 'scsi-fixes' of git://git./linux/kernel/git/jejb/scsi

Pull SCSI fix from James Bottomley:
 "Just a single lpfc fix adjusting the number of available queues for
  high CPU count systems"

* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
  scsi: lpfc: Raise config max for lpfc_fcp_mq_threshold variable

5 years agoMerge tag 'libnvdimm-fix-5.3-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 6 Sep 2019 23:14:32 +0000 (16:14 -0700)]
Merge tag 'libnvdimm-fix-5.3-rc8' of git://git./linux/kernel/git/nvdimm/nvdimm

Pull libnvdimm fix from Dan Williams:
 "Restore support for 1GB alignment namespaces, truncate the end of
  misaligned namespaces"

* tag 'libnvdimm-fix-5.3-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm:
  libnvdimm/pfn: Fix namespace creation on misaligned addresses

5 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input
Linus Torvalds [Fri, 6 Sep 2019 23:12:30 +0000 (16:12 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/dtor/input

Pull input fix from Dmitry Torokhov:
 "A tiny update from Benjamin removing a mistakenly added Elan PNP ID so
  that the device is again handled by hid-multitouch"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
  Input: elan_i2c - remove Lenovo Legion Y7000 PnpID

5 years agoInput: elan_i2c - remove Lenovo Legion Y7000 PnpID
Benjamin Tissoires [Fri, 6 Sep 2019 19:02:32 +0000 (12:02 -0700)]
Input: elan_i2c - remove Lenovo Legion Y7000 PnpID

Looks like the Bios of the Lenovo Legion Y7000 is using ELAN061B
when the actual device is supposed to be used with hid-multitouch.

Remove it from the list of the supported device, hoping that
no one will complain about the loss in functionality.

Link: https://bugzilla.kernel.org/show_bug.cgi?id=203467
Fixes: 738c06d0e456 ("Input: elan_i2c - add hardware ID for multiple Lenovo laptops")
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
5 years agoMerge tag 'armsoc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc
Linus Torvalds [Fri, 6 Sep 2019 19:53:31 +0000 (12:53 -0700)]
Merge tag 'armsoc-fixes' of git://git./linux/kernel/git/soc/soc

Pull ARM SoC fixes from Arnd Bergmann:
 "There are three more fixes for this week:

   - The Windows-on-ARM laptops require a workaround to prevent crashing
     at boot from ACPI

   - The Renesas 'draak' board needs one bugfix for the backlight
     regulator

   - Also for Renesas, the 'hihope' board accidentally had its eMMC
     turned off in the 5.3 merge window"

* tag 'armsoc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc:
  soc: qcom: geni: Provide parameter error checking
  arm64: dts: renesas: hihope-common: Fix eMMC status
  arm64: dts: renesas: r8a77995: draak: Fix backlight regulator name

5 years agoMerge tag 'configfs-for-5.3' of git://git.infradead.org/users/hch/configfs
Linus Torvalds [Fri, 6 Sep 2019 19:44:08 +0000 (12:44 -0700)]
Merge tag 'configfs-for-5.3' of git://git.infradead.org/users/hch/configfs

Pull configfs fixes from Christoph Hellwig:
 "Late configfs fixes from Al that fix pretty nasty removal vs attribute
  access races"

* tag 'configfs-for-5.3' of git://git.infradead.org/users/hch/configfs:
  configfs: provide exclusion between IO and removals
  configfs: new object reprsenting tree fragments
  configfs_register_group() shouldn't be (and isn't) called in rmdirable parts
  configfs: stash the data we need into configfs_buffer at open time

5 years agoMerge tag 'iommu-fixes-v5.3-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 6 Sep 2019 19:22:36 +0000 (12:22 -0700)]
Merge tag 'iommu-fixes-v5.3-rc7' of git://git./linux/kernel/git/joro/iommu

Pull IOMMU fixes from Joerg Roedel:

 - Revert an Intel VT-d patch that caused problems for some users.

 - Removal of a feature in the Intel VT-d driver that was never
   supported in hardware. This qualifies as a fix because the code for
   this feature sets reserved bits in the invalidation queue descriptor,
   causing failed invalidations on real hardware.

 - Two fixes for AMD IOMMU driver to fix a race condition and to add a
   missing IOTLB flush when kernel is booted in kdump mode.

* tag 'iommu-fixes-v5.3-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu:
  iommu/amd: Fix race in increase_address_space()
  iommu/amd: Flush old domains in kdump kernel
  iommu/vt-d: Remove global page flush support
  Revert "iommu/vt-d: Avoid duplicated pci dma alias consideration"

5 years agoMerge tag 'mmc-v5.3-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc
Linus Torvalds [Fri, 6 Sep 2019 16:01:49 +0000 (09:01 -0700)]
Merge tag 'mmc-v5.3-rc7' of git://git./linux/kernel/git/ulfh/mmc

Pull MMC fix from Ulf Hansson:
 "Revert in order to fix card init for some eMMCs that need retries for
  CMD6"

* tag 'mmc-v5.3-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc:
  Revert "mmc: core: do not retry CMD6 in __mmc_switch()"

5 years agoMerge tag 'drm-fixes-2019-09-06' of git://anongit.freedesktop.org/drm/drm
Linus Torvalds [Fri, 6 Sep 2019 15:58:19 +0000 (08:58 -0700)]
Merge tag 'drm-fixes-2019-09-06' of git://anongit.freedesktop.org/drm/drm

Pull drm fixes from Dave Airlie:
 "Live from my friend's couch in Barcelona, latest round of drm fixes.

  The command line parser regression fixes look a bit larger because
  they come with selftests included for the bugs they fix. Otherwise a
  single nouveau, single ingenic and single vmwgfx fix:

  nouveau:
   - add missing MODULE_FIRMWARE definitions

  igenic:
   - hardcode panel type DPI

  vmwgfx:
   - double free fix

  core:
   - command line mode parser fixes"

* tag 'drm-fixes-2019-09-06' of git://anongit.freedesktop.org/drm/drm:
  drm/vmwgfx: Fix double free in vmw_recv_msg()
  drm/nouveau/sec2/gp102: add missing MODULE_FIRMWAREs
  drm/selftests: modes: Add more unit tests for the cmdline parser
  drm/modes: Introduce a whitelist for the named modes
  drm/modes: Fix the command line parser to take force options into account
  drm/modes: Add a switch to differentiate free standing options
  drm/ingenic: Hardcode panel type to DPI

5 years agoMerge tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost
Linus Torvalds [Fri, 6 Sep 2019 15:56:06 +0000 (08:56 -0700)]
Merge tag 'for_linus' of git://git./linux/kernel/git/mst/vhost

Pull virtio fixes from Michael Tsirkin:
 "virtio, vhost, and balloon bugfixes.

  A couple of last minute bugfixes. And a revert of a failed attempt at
  metadata access optimization - we'll try again in the next cycle"

* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost:
  mm/balloon_compaction: suppress allocation warnings
  Revert "vhost: access vq metadata through kernel virtual address"
  vhost: Remove unnecessary variable
  virtio-net: lower min ring num_free for efficiency
  vhost/test: fix build for vhost test
  vhost/test: fix build for vhost test

5 years agoMerge tag 'powerpc-5.3-5' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc...
Linus Torvalds [Fri, 6 Sep 2019 15:54:45 +0000 (08:54 -0700)]
Merge tag 'powerpc-5.3-5' of git://git./linux/kernel/git/powerpc/linux

Pull powerpc fixes from Michael Ellerman:
 "One fix for a boot hang on some Freescale machines when PREEMPT is
  enabled.

  Two CVE fixes for bugs in our handling of FP registers and
  transactional memory, both of which can result in corrupted FP state,
  or FP state leaking between processes.

  Thanks to: Chris Packham, Christophe Leroy, Gustavo Romero, Michael
  Neuling"

* tag 'powerpc-5.3-5' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
  powerpc/tm: Fix restoring FP/VMX facility incorrectly on interrupts
  powerpc/tm: Fix FP/VMX unavailable exceptions inside a transaction
  powerpc/64e: Drop stale call to smp_processor_id() which hangs SMP startup

5 years agoDocumentation/process/embargoed-hardware-issues: Microsoft ambassador
Sasha Levin [Fri, 6 Sep 2019 09:58:52 +0000 (05:58 -0400)]
Documentation/process/embargoed-hardware-issues: Microsoft ambassador

Add Sasha Levin as Microsoft's process ambassador.

Signed-off-by: Sasha Levin <alexander.levin@microsoft.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
Link: https://lore.kernel.org/r/20190906095852.23568-1-sashal@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agosoc: qcom: geni: Provide parameter error checking
Lee Jones [Thu, 5 Sep 2019 08:25:55 +0000 (09:25 +0100)]
soc: qcom: geni: Provide parameter error checking

When booting with ACPI, the Geni Serial Engine is not set as the I2C/SPI
parent and thus, the wrapper (parent device) is unassigned.  This causes
the kernel to crash with a null dereference error.

Link: https://lore.kernel.org/r/20190905082555.15020-1-lee.jones@linaro.org
Fixes: 8bc529b25354 ("soc: qcom: geni: Add support for ACPI")
Acked-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Reviewed-by: Stephen Boyd <sboyd@kernel.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
5 years agoiommu/amd: Fix race in increase_address_space()
Joerg Roedel [Fri, 6 Sep 2019 08:39:54 +0000 (10:39 +0200)]
iommu/amd: Fix race in increase_address_space()

After the conversion to lock-less dma-api call the
increase_address_space() function can be called without any
locking. Multiple CPUs could potentially race for increasing
the address space, leading to invalid domain->mode settings
and invalid page-tables. This has been happening in the wild
under high IO load and memory pressure.

Fix the race by locking this operation. The function is
called infrequently so that this does not introduce
a performance regression in the dma-api path again.

Reported-by: Qian Cai <cai@lca.pw>
Fixes: 256e4621c21a ('iommu/amd: Make use of the generic IOVA allocator')
Signed-off-by: Joerg Roedel <jroedel@suse.de>
5 years agoiommu/amd: Flush old domains in kdump kernel
Stuart Hayes [Thu, 5 Sep 2019 17:09:48 +0000 (12:09 -0500)]
iommu/amd: Flush old domains in kdump kernel

When devices are attached to the amd_iommu in a kdump kernel, the old device
table entries (DTEs), which were copied from the crashed kernel, will be
overwritten with a new domain number.  When the new DTE is written, the IOMMU
is told to flush the DTE from its internal cache--but it is not told to flush
the translation cache entries for the old domain number.

Without this patch, AMD systems using the tg3 network driver fail when kdump
tries to save the vmcore to a network system, showing network timeouts and
(sometimes) IOMMU errors in the kernel log.

This patch will flush IOMMU translation cache entries for the old domain when
a DTE gets overwritten with a new domain number.

Signed-off-by: Stuart Hayes <stuart.w.hayes@gmail.com>
Fixes: 3ac3e5ee5ed5 ('iommu/amd: Copy old trans table from old kernel')
Signed-off-by: Joerg Roedel <jroedel@suse.de>
5 years agoMerge tag 'drm-misc-fixes-2019-09-05' of git://anongit.freedesktop.org/drm/drm-misc...
Dave Airlie [Fri, 6 Sep 2019 06:25:45 +0000 (16:25 +1000)]
Merge tag 'drm-misc-fixes-2019-09-05' of git://anongit.freedesktop.org/drm/drm-misc into drm-fixes

drm-misc-fixes for v5.3 final:
- Make ingenic panel type DPI insteado f unknown.
- Fixes for command line parser modes.

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/606d87b2-1840-c893-eb30-d6c471c9e50a@linux.intel.com
5 years agoMerge branch 'vmwgfx-fixes-5.3' of git://people.freedesktop.org/~thomash/linux into...
Dave Airlie [Fri, 6 Sep 2019 06:23:25 +0000 (16:23 +1000)]
Merge branch 'vmwgfx-fixes-5.3' of git://people.freedesktop.org/~thomash/linux into drm-fixes

Single vmwgfx double free fix.

Signed-off-by: Dave Airlie <airlied@redhat.com>
5 years agokeys: Fix missing null pointer check in request_key_auth_describe()
Hillf Danton [Mon, 2 Sep 2019 12:37:29 +0000 (13:37 +0100)]
keys: Fix missing null pointer check in request_key_auth_describe()

If a request_key authentication token key gets revoked, there's a window in
which request_key_auth_describe() can see it with a NULL payload - but it
makes no check for this and something like the following oops may occur:

BUG: Kernel NULL pointer dereference at 0x00000038
Faulting instruction address: 0xc0000000004ddf30
Oops: Kernel access of bad area, sig: 11 [#1]
...
NIP [...] request_key_auth_describe+0x90/0xd0
LR [...] request_key_auth_describe+0x54/0xd0
Call Trace:
[...] request_key_auth_describe+0x54/0xd0 (unreliable)
[...] proc_keys_show+0x308/0x4c0
[...] seq_read+0x3d0/0x540
[...] proc_reg_read+0x90/0x110
[...] __vfs_read+0x3c/0x70
[...] vfs_read+0xb4/0x1b0
[...] ksys_read+0x7c/0x130
[...] system_call+0x5c/0x70

Fix this by checking for a NULL pointer when describing such a key.

Also make the read routine check for a NULL pointer to be on the safe side.

[DH: Modified to not take already-held rcu lock and modified to also check
 in the read routine]

Fixes: 04c567d9313e ("[PATCH] Keys: Fix race between two instantiators of a key")
Reported-by: Sachin Sant <sachinp@linux.vnet.ibm.com>
Signed-off-by: Hillf Danton <hdanton@sina.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Tested-by: Sachin Sant <sachinp@linux.vnet.ibm.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agoMerge tag 'sound-5.3-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai...
Linus Torvalds [Thu, 5 Sep 2019 17:26:20 +0000 (10:26 -0700)]
Merge tag 'sound-5.3-rc8' of git://git./linux/kernel/git/tiwai/sound

Pull sound fixes from Takashi Iwai:
 "A collection of small HD-audio fixes:

   - A regression fix for Realtek codecs due to the recent
     initialization procedure change

   - A fix for potential endless loop at the quirk table lookup

   - Quirks for Lenovo, ASUS and HP machines"

* tag 'sound-5.3-rc8' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
  ALSA: hda/realtek - Fix the problem of two front mics on a ThinkCentre
  ALSA: hda/realtek - Enable internal speaker & headset mic of ASUS UX431FL
  ALSA: hda/realtek - Add quirk for HP Pavilion 15
  ALSA: hda/realtek - Fix overridden device-specific initialization
  ALSA: hda - Fix potential endless loop at applying quirks

5 years agoASoC: SOF: imx8: Fix COMPILE_TEST error
YueHaibing [Thu, 5 Sep 2019 06:44:00 +0000 (14:44 +0800)]
ASoC: SOF: imx8: Fix COMPILE_TEST error

When do compile test, if SND_SOC_SOF_OF is not set, we get:

sound/soc/sof/imx/imx8.o: In function `imx8_dsp_handle_request':
imx8.c:(.text+0xb0): undefined reference to `snd_sof_ipc_msgs_rx'
sound/soc/sof/imx/imx8.o: In function `imx8_ipc_msg_data':
imx8.c:(.text+0xf4): undefined reference to `sof_mailbox_read'
sound/soc/sof/imx/imx8.o: In function `imx8_dsp_handle_reply':
imx8.c:(.text+0x160): undefined reference to `sof_mailbox_read'

Make SND_SOC_SOF_IMX_TOPLEVEL always depends on SND_SOC_SOF_OF

Reported-by: Hulk Robot <hulkci@huawei.com>
Fixes: 202acc565a1f ("ASoC: SOF: imx: Add i.MX8 HW support")
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Reviewed-by: Daniel Baluta <daniel.baluta@nxp.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20190905064400.24800-1-yuehaibing@huawei.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: meson: tdmout: add sm1 support
Jerome Brunet [Thu, 5 Sep 2019 12:01:20 +0000 (14:01 +0200)]
ASoC: meson: tdmout: add sm1 support

On the sm1, the TDMOUT number of input is extended and the
the gain enable bit moved to accommodate this extension

Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Link: https://lore.kernel.org/r/20190905120120.31752-9-jbrunet@baylibre.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: meson: axg-toddr: add sm1 support
Jerome Brunet [Thu, 5 Sep 2019 12:01:19 +0000 (14:01 +0200)]
ASoC: meson: axg-toddr: add sm1 support

On sm1, the maximum number TODDR inputs is extended to 16.

Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Link: https://lore.kernel.org/r/20190905120120.31752-8-jbrunet@baylibre.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: meson: axg-frddr: add sm1 support
Jerome Brunet [Thu, 5 Sep 2019 12:01:18 +0000 (14:01 +0200)]
ASoC: meson: axg-frddr: add sm1 support

On sm1, the output routing bits have moved to CTRL2 register

Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Link: https://lore.kernel.org/r/20190905120120.31752-7-jbrunet@baylibre.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: meson: tdmin: expose all 16 inputs
Jerome Brunet [Thu, 5 Sep 2019 12:01:17 +0000 (14:01 +0200)]
ASoC: meson: tdmin: expose all 16 inputs

The TDMIN component, as it, has a maximum of 16 input. Depending on
the SoC, these may not all be connected.

Instead of decribing only the connected inputs of each SoC, describe
them all and let ASoC routing do the rest.

Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Link: https://lore.kernel.org/r/20190905120120.31752-6-jbrunet@baylibre.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: meson: axg-toddr: expose all 8 inputs
Jerome Brunet [Thu, 5 Sep 2019 12:01:16 +0000 (14:01 +0200)]
ASoC: meson: axg-toddr: expose all 8 inputs

The TODDR component, as it, has a maximum of 8 input. Depending on
the SoC, these may not all be connected or some input components may
not be supported

Instead of decribing only the connected inputs, describe them all
and let ASoC routing do the rest.

Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Link: https://lore.kernel.org/r/20190905120120.31752-5-jbrunet@baylibre.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: meson: axg-frddr: expose all 8 outputs
Jerome Brunet [Thu, 5 Sep 2019 12:01:15 +0000 (14:01 +0200)]
ASoC: meson: axg-frddr: expose all 8 outputs

The FRDDR component, as it, has a maximum of 8 outputs. Depending on
the SoC, these may not all be connected.

Instead of decribing only the connected outputs of each SoC, describe
them all and let ASoC routing do the rest.

Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Link: https://lore.kernel.org/r/20190905120120.31752-4-jbrunet@baylibre.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: meson: add reset binding
Jerome Brunet [Thu, 5 Sep 2019 12:01:14 +0000 (14:01 +0200)]
ASoC: meson: add reset binding

The g12a audio subsystem, which is a derivative of the axg subsystem,
provides a dedicated reset line for each of the audio components.

The axg did not provide that and it is unclear if/when these reset are
required. The reset already helped solve a channel mapping issue on the
tdm formatter devices. Let's add the reset binding for the other
components, so we can describe this in DT. We'll use it later on
in the driver when/if needed.

Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Link: https://lore.kernel.org/r/20190905120120.31752-3-jbrunet@baylibre.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: meson: add sm1 compatibles
Jerome Brunet [Thu, 5 Sep 2019 12:01:13 +0000 (14:01 +0200)]
ASoC: meson: add sm1 compatibles

Document the compatible strings of the audio devices of the sm1 SoC
family

Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Link: https://lore.kernel.org/r/20190905120120.31752-2-jbrunet@baylibre.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: qcom: common: Include link-name in error messages
Bjorn Andersson [Thu, 5 Sep 2019 04:03:06 +0000 (21:03 -0700)]
ASoC: qcom: common: Include link-name in error messages

Reading out the link-name earlier and including it in the various error
messages makes it much more convenient to figure out what links have
unmet dependencies.

Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20190905040306.21399-1-bjorn.andersson@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoMerge branch 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Thu, 5 Sep 2019 16:47:32 +0000 (09:47 -0700)]
Merge branch 'x86-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull x86 fixes from Ingo Molnar:
 "Misc fixes:

   - EFI boot fix for signed kernels

   - an AC flags fix related to UBSAN

   - Hyper-V infinite loop fix"

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/hyper-v: Fix overflow bug in fill_gva_list()
  x86/uaccess: Don't leak the AC flags into __get_user() argument evaluation
  x86/boot: Preserve boot_params.secure_boot from sanitizing

5 years agoMerge branch 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Thu, 5 Sep 2019 16:38:31 +0000 (09:38 -0700)]
Merge branch 'sched-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull scheduler fixes from Ingo Molnar:
 "This fixes an ABI bug introduced this cycle, plus fixes a throttling
  bug"

* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched/core: Fix uclamp ABI bug, clean up and robustify sched_read_attr() ABI logic and code
  sched/fair: Don't assign runtime for throttled cfs_rq

5 years agoMerge tag 'clang-format-for-linus-v5.3-rc8' of git://github.com/ojeda/linux
Linus Torvalds [Thu, 5 Sep 2019 16:23:02 +0000 (09:23 -0700)]
Merge tag 'clang-format-for-linus-v5.3-rc8' of git://github.com/ojeda/linux

Pull clang-format update from Miguel Ojeda:
 "Update with the latest for_each macro list"

* tag 'clang-format-for-linus-v5.3-rc8' of git://github.com/ojeda/linux:
  clang-format: Update with the latest for_each macro list

5 years agoMerge tag 'renesas-fixes2-for-v5.3' of git://git.kernel.org/pub/scm/linux/kernel...
Arnd Bergmann [Thu, 5 Sep 2019 15:56:29 +0000 (17:56 +0200)]
Merge tag 'renesas-fixes2-for-v5.3' of git://git./linux/kernel/git/horms/renesas into arm/fixes

Second Round of Renesas ARM Based SoC Fixes for v5.3

* RZ/G2M based HiHope main board
  - Re-enabled accidently disabled SDHI3 (eMMC) support

* tag 'renesas-fixes2-for-v5.3' of git://git.kernel.org/pub/scm/linux/kernel/git/horms/renesas:
  arm64: dts: renesas: hihope-common: Fix eMMC status

Link: https://lore.kernel.org/r/cover.1567675986.git.horms+renesas@verge.net.au
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
5 years agodrm/vmwgfx: Fix double free in vmw_recv_msg()
Dan Carpenter [Thu, 15 Aug 2019 08:30:50 +0000 (11:30 +0300)]
drm/vmwgfx: Fix double free in vmw_recv_msg()

We recently added a kfree() after the end of the loop:

if (retries == RETRIES) {
kfree(reply);
return -EINVAL;
}

There are two problems.  First the test is wrong and because retries
equals RETRIES if we succeed on the last iteration through the loop.
Second if we fail on the last iteration through the loop then the kfree
is a double free.

When you're reading this code, please note the break statement at the
end of the while loop.  This patch changes the loop so that if it's not
successful then "reply" is NULL and we can test for that afterward.

Cc: <stable@vger.kernel.org>
Fixes: 6b7c3b86f0b6 ("drm/vmwgfx: fix memory leak when too many retries have occurred")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
5 years agoDocumentation/process: Add Google contact for embargoed hardware issues
Kees Cook [Wed, 4 Sep 2019 16:24:49 +0000 (09:24 -0700)]
Documentation/process: Add Google contact for embargoed hardware issues

This adds myself as the Google contact for embargoed hardware security
issues and fixes some small typos.

Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Matt Linton <amuse@google.com>
Cc: Matthew Garrett <mjg59@google.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Acked-by: Guenter Roeck <groeck@chromium.org>
Link: https://lore.kernel.org/r/201909040922.56496BF70@keescook
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agoDocumentation/process: Volunteer as the ambassador for Xen
Andrew Cooper [Wed, 4 Sep 2019 18:17:02 +0000 (19:17 +0100)]
Documentation/process: Volunteer as the ambassador for Xen

Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tyler Hicks <tyhicks@canonical.com>
Cc: Ben Hutchings <ben@decadent.org.uk>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Link: https://lore.kernel.org/r/20190904181702.19788-1-andrew.cooper3@citrix.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agoconfigfs: provide exclusion between IO and removals
Al Viro [Sat, 31 Aug 2019 07:43:43 +0000 (09:43 +0200)]
configfs: provide exclusion between IO and removals

Make sure that attribute methods are not called after the item
has been removed from the tree.  To do so, we
* at the point of no return in removals, grab ->frag_sem
exclusive and mark the fragment dead.
* call the methods of attributes with ->frag_sem taken
shared and only after having verified that the fragment is still
alive.

The main benefit is for method instances - they are
guaranteed that the objects they are accessing *and* all ancestors
are still there.  Another win is that we don't need to bother
with extra refcount on config_item when opening a file -
the item will be alive for as long as it stays in the tree, and
we won't touch it/attributes/any associated data after it's
been removed from the tree.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Christoph Hellwig <hch@lst.de>
5 years agoMerge tag 'nfs-for-5.3-4' of git://git.linux-nfs.org/projects/trondmy/linux-nfs
Linus Torvalds [Wed, 4 Sep 2019 18:09:53 +0000 (11:09 -0700)]
Merge tag 'nfs-for-5.3-4' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

Pull NFS client bugfix from Trond Myklebust:
 "Regression fix inode fileid checks in attribute revalidation code"

* tag 'nfs-for-5.3-4' of git://git.linux-nfs.org/projects/trondmy/linux-nfs:
  NFS: Fix inode fileid checks in attribute revalidation code

5 years agosched/core: Fix uclamp ABI bug, clean up and robustify sched_read_attr() ABI logic...
Ingo Molnar [Wed, 4 Sep 2019 07:55:32 +0000 (09:55 +0200)]
sched/core: Fix uclamp ABI bug, clean up and robustify sched_read_attr() ABI logic and code

Thadeu Lima de Souza Cascardo reported that 'chrt' broke on recent kernels:

  $ chrt -p $$
  chrt: failed to get pid 26306's policy: Argument list too long

and he has root-caused the bug to the following commit increasing sched_attr
size and breaking sched_read_attr() into returning -EFBIG:

  a509a7cd7974 ("sched/uclamp: Extend sched_setattr() to support utilization clamping")

The other, bigger bug is that the whole sched_getattr() and sched_read_attr()
logic of checking non-zero bits in new ABI components is arguably broken,
and pretty much any extension of the ABI will spuriously break the ABI.
That's way too fragile.

Instead implement the perf syscall's extensible ABI instead, which we
already implement on the sched_setattr() side:

 - if user-attributes have the same size as kernel attributes then the
   logic is unchanged.

 - if user-attributes are larger than the kernel knows about then simply
   skip the extra bits, but set attr->size to the (smaller) kernel size
   so that tooling can (in principle) handle older kernel as well.

 - if user-attributes are smaller than the kernel knows about then just
   copy whatever user-space can accept.

Also clean up the whole logic:

 - Simplify the code flow - there's no need for 'ret' for example.

 - Standardize on 'kattr/uattr' and 'ksize/usize' naming to make sure we
   always know which side we are dealing with.

 - Why is it called 'read' when what it does is to copy to user? This
   code is so far away from VFS read() semantics that the naming is
   actively confusing. Name it sched_attr_copy_to_user() instead, which
   mirrors other copy_to_user() functionality.

 - Move the attr->size assignment from the head of sched_getattr() to the
   sched_attr_copy_to_user() function. Nothing else within the kernel
   should care about the size of the structure.

With these fixes the sched_getattr() syscall now nicely supports an
extensible ABI in both a forward and backward compatible fashion, and
will also fix the chrt bug.

As an added bonus the bogus -EFBIG return is removed as well, which as
Thadeu noted should have been -E2BIG to begin with.

Reported-by: Thadeu Lima de Souza Cascardo <cascardo@canonical.com>
Tested-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Tested-by: Thadeu Lima de Souza Cascardo <cascardo@canonical.com>
Acked-by: Thadeu Lima de Souza Cascardo <cascardo@canonical.com>
Cc: Arnaldo Carvalho de Melo <acme@infradead.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Patrick Bellasi <patrick.bellasi@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Fixes: a509a7cd7974 ("sched/uclamp: Extend sched_setattr() to support utilization clamping")
Link: https://lkml.kernel.org/r/20190904075532.GA26751@gmail.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
5 years agoASoC: soc-core: add comment to jack at soc_remove_component()
Kuninori Morimoto [Wed, 4 Sep 2019 00:14:35 +0000 (09:14 +0900)]
ASoC: soc-core: add comment to jack at soc_remove_component()

Basically, driver which setups snd_soc_component_set_jack() need
to release it by themselves. But, as framework level robustness,
soc_remove_component() also releases it.

To avoid code reader confuse, this patch makes it clarify.

This patch makes it clarify.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/875zm8q5n8.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: tegra: use devm_platform_ioremap_resource() to simplify code
YueHaibing [Wed, 4 Sep 2019 08:39:09 +0000 (16:39 +0800)]
ASoC: tegra: use devm_platform_ioremap_resource() to simplify code

Use devm_platform_ioremap_resource() to simplify the code a bit.
This is detected by coccinelle.

Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Link: https://lore.kernel.org/r/20190904083909.18804-1-yuehaibing@huawei.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: sirf-audio: use devm_platform_ioremap_resource() to simplify code
YueHaibing [Wed, 4 Sep 2019 08:34:12 +0000 (16:34 +0800)]
ASoC: sirf-audio: use devm_platform_ioremap_resource() to simplify code

Use devm_platform_ioremap_resource() to simplify the code a bit.
This is detected by coccinelle.

Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Link: https://lore.kernel.org/r/20190904083412.18700-1-yuehaibing@huawei.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: ep93xx: use devm_platform_ioremap_resource() to simplify code
YueHaibing [Wed, 4 Sep 2019 08:25:07 +0000 (16:25 +0800)]
ASoC: ep93xx: use devm_platform_ioremap_resource() to simplify code

Use devm_platform_ioremap_resource() to simplify the code a bit.
This is detected by coccinelle.

Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Link: https://lore.kernel.org/r/20190904082507.24300-1-yuehaibing@huawei.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: amd: use devm_platform_ioremap_resource() to simplify code
YueHaibing [Wed, 4 Sep 2019 07:48:33 +0000 (15:48 +0800)]
ASoC: amd: use devm_platform_ioremap_resource() to simplify code

Use devm_platform_ioremap_resource() to simplify the code a bit.
This is detected by coccinelle.

Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Link: https://lore.kernel.org/r/20190904074833.23572-1-yuehaibing@huawei.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoMerge tag 'renesas-fixes-for-v5.3' of git://git.kernel.org/pub/scm/linux/kernel/git...
Arnd Bergmann [Wed, 4 Sep 2019 12:38:34 +0000 (14:38 +0200)]
Merge tag 'renesas-fixes-for-v5.3' of git://git./linux/kernel/git/horms/renesas into arm/fixes

Renesas ARM Based SoC Fixes for v5.3

* R-Car D3 (r8a77995) based Draak Board
  - Correct backlight regulator name in device tree

* tag 'renesas-fixes-for-v5.3' of git://git.kernel.org/pub/scm/linux/kernel/git/horms/renesas:
  arm64: dts: renesas: r8a77995: draak: Fix backlight regulator name

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
5 years agopowerpc/tm: Fix restoring FP/VMX facility incorrectly on interrupts
Gustavo Romero [Wed, 4 Sep 2019 04:55:28 +0000 (00:55 -0400)]
powerpc/tm: Fix restoring FP/VMX facility incorrectly on interrupts

When in userspace and MSR FP=0 the hardware FP state is unrelated to
the current process. This is extended for transactions where if tbegin
is run with FP=0, the hardware checkpoint FP state will also be
unrelated to the current process. Due to this, we need to ensure this
hardware checkpoint is updated with the correct state before we enable
FP for this process.

Unfortunately we get this wrong when returning to a process from a
hardware interrupt. A process that starts a transaction with FP=0 can
take an interrupt. When the kernel returns back to that process, we
change to FP=1 but with hardware checkpoint FP state not updated. If
this transaction is then rolled back, the FP registers now contain the
wrong state.

The process looks like this:
   Userspace:                      Kernel

               Start userspace
                with MSR FP=0 TM=1
                  < -----
   ...
   tbegin
   bne
               Hardware interrupt
                   ---- >
                                    <do_IRQ...>
                                    ....
                                    ret_from_except
                                      restore_math()
        /* sees FP=0 */
                                        restore_fp()
                                          tm_active_with_fp()
    /* sees FP=1 (Incorrect) */
                                          load_fp_state()
                                        FP = 0 -> 1
                  < -----
               Return to userspace
                 with MSR TM=1 FP=1
                 with junk in the FP TM checkpoint
   TM rollback
   reads FP junk

When returning from the hardware exception, tm_active_with_fp() is
incorrectly making restore_fp() call load_fp_state() which is setting
FP=1.

The fix is to remove tm_active_with_fp().

tm_active_with_fp() is attempting to handle the case where FP state
has been changed inside a transaction. In this case the checkpointed
and transactional FP state is different and hence we must restore the
FP state (ie. we can't do lazy FP restore inside a transaction that's
used FP). It's safe to remove tm_active_with_fp() as this case is
handled by restore_tm_state(). restore_tm_state() detects if FP has
been using inside a transaction and will set load_fp and call
restore_math() to ensure the FP state (checkpoint and transaction) is
restored.

This is a data integrity problem for the current process as the FP
registers are corrupted. It's also a security problem as the FP
registers from one process may be leaked to another.

Similarly for VMX.

A simple testcase to replicate this will be posted to
tools/testing/selftests/powerpc/tm/tm-poison.c

This fixes CVE-2019-15031.

Fixes: a7771176b439 ("powerpc: Don't enable FP/Altivec if not checkpointed")
Cc: stable@vger.kernel.org # 4.15+
Signed-off-by: Gustavo Romero <gromero@linux.ibm.com>
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190904045529.23002-2-gromero@linux.vnet.ibm.com
5 years agopowerpc/tm: Fix FP/VMX unavailable exceptions inside a transaction
Gustavo Romero [Wed, 4 Sep 2019 04:55:27 +0000 (00:55 -0400)]
powerpc/tm: Fix FP/VMX unavailable exceptions inside a transaction

When we take an FP unavailable exception in a transaction we have to
account for the hardware FP TM checkpointed registers being
incorrect. In this case for this process we know the current and
checkpointed FP registers must be the same (since FP wasn't used
inside the transaction) hence in the thread_struct we copy the current
FP registers to the checkpointed ones.

This copy is done in tm_reclaim_thread(). We use thread->ckpt_regs.msr
to determine if FP was on when in userspace. thread->ckpt_regs.msr
represents the state of the MSR when exiting userspace. This is setup
by check_if_tm_restore_required().

Unfortunatley there is an optimisation in giveup_all() which returns
early if tsk->thread.regs->msr (via local variable `usermsr`) has
FP=VEC=VSX=SPE=0. This optimisation means that
check_if_tm_restore_required() is not called and hence
thread->ckpt_regs.msr is not updated and will contain an old value.

This can happen if due to load_fp=255 we start a userspace process
with MSR FP=1 and then we are context switched out. In this case
thread->ckpt_regs.msr will contain FP=1. If that same process is then
context switched in and load_fp overflows, MSR will have FP=0. If that
process now enters a transaction and does an FP instruction, the FP
unavailable will not update thread->ckpt_regs.msr (the bug) and MSR
FP=1 will be retained in thread->ckpt_regs.msr.  tm_reclaim_thread()
will then not perform the required memcpy and the checkpointed FP regs
in the thread struct will contain the wrong values.

The code path for this happening is:

       Userspace:                      Kernel
                   Start userspace
                    with MSR FP/VEC/VSX/SPE=0 TM=1
                      < -----
       ...
       tbegin
       bne
       fp instruction
                   FP unavailable
                       ---- >
                                        fp_unavailable_tm()
  tm_reclaim_current()
    tm_reclaim_thread()
      giveup_all()
        return early since FP/VMX/VSX=0
/* ckpt MSR not updated (Incorrect) */
      tm_reclaim()
        /* thread_struct ckpt FP regs contain junk (OK) */
                                              /* Sees ckpt MSR FP=1 (Incorrect) */
      no memcpy() performed
        /* thread_struct ckpt FP regs not fixed (Incorrect) */
  tm_recheckpoint()
     /* Put junk in hardware checkpoint FP regs */
                                         ....
                      < -----
                   Return to userspace
                     with MSR TM=1 FP=1
                     with junk in the FP TM checkpoint
       TM rollback
       reads FP junk

This is a data integrity problem for the current process as the FP
registers are corrupted. It's also a security problem as the FP
registers from one process may be leaked to another.

This patch moves up check_if_tm_restore_required() in giveup_all() to
ensure thread->ckpt_regs.msr is updated correctly.

A simple testcase to replicate this will be posted to
tools/testing/selftests/powerpc/tm/tm-poison.c

Similarly for VMX.

This fixes CVE-2019-15030.

Fixes: f48e91e87e67 ("powerpc/tm: Fix FP and VMX register corruption")
Cc: stable@vger.kernel.org # 4.12+
Signed-off-by: Gustavo Romero <gromero@linux.vnet.ibm.com>
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190904045529.23002-1-gromero@linux.vnet.ibm.com
5 years agomm/balloon_compaction: suppress allocation warnings
Nadav Amit [Tue, 20 Aug 2019 09:16:46 +0000 (02:16 -0700)]
mm/balloon_compaction: suppress allocation warnings

There is no reason to print warnings when balloon page allocation fails,
as they are expected and can be handled gracefully.  Since VMware
balloon now uses balloon-compaction infrastructure, and suppressed these
warnings before, it is also beneficial to suppress these warnings to
keep the same behavior that the balloon had before.

Cc: Jason Wang <jasowang@redhat.com>
Signed-off-by: Nadav Amit <namit@vmware.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
5 years agoRevert "vhost: access vq metadata through kernel virtual address"
Michael S. Tsirkin [Sat, 10 Aug 2019 17:53:21 +0000 (13:53 -0400)]
Revert "vhost: access vq metadata through kernel virtual address"

This reverts commit 7f466032dc ("vhost: access vq metadata through
kernel virtual address").  The commit caused a bunch of issues, and
while commit 73f628ec9e ("vhost: disable metadata prefetch
optimization") disabled the optimization it's not nice to keep lots of
dead code around.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
5 years agovhost: Remove unnecessary variable
Yunsheng Lin [Tue, 20 Aug 2019 12:36:32 +0000 (20:36 +0800)]
vhost: Remove unnecessary variable

It is unnecessary to use ret variable to return the error
code, just return the error code directly.

Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
5 years agovirtio-net: lower min ring num_free for efficiency
? jiang [Tue, 20 Aug 2019 02:51:23 +0000 (02:51 +0000)]
virtio-net: lower min ring num_free for efficiency

This change lowers ring buffer reclaim threshold from 1/2*queue to budget
for better performance. According to our test with qemu + dpdk, packet
dropping happens when the guest is not able to provide free buffer in
avail ring timely with default 1/2*queue. The value in the patch has been
tested and does show better performance.

Test setup: iperf3 to generate packets to guest (total 30mins, pps 400k, UDP)
avg packets drop before: 2842
avg packets drop after: 360(-87.3%)

Further, current code suffers from a starvation problem: the amount of
work done by try_fill_recv is not bounded by the budget parameter, thus
(with large queues) once in a while userspace gets blocked for a long
time while queue is being refilled. Trigger refills earlier to make sure
the amount of work to do is limited.

Signed-off-by: jiangkidd <jiangkidd@hotmail.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
5 years agovhost/test: fix build for vhost test
Tiwei Bie [Wed, 28 Aug 2019 05:37:00 +0000 (13:37 +0800)]
vhost/test: fix build for vhost test

Since vhost_exceeds_weight() was introduced, callers need to specify
the packet weight and byte weight in vhost_dev_init(). Note that, the
packet weight isn't counted in this patch to keep the original behavior
unchanged.

Fixes: e82b9b0727ff ("vhost: introduce vhost_exceeds_weight()")
Cc: stable@vger.kernel.org
Signed-off-by: Tiwei Bie <tiwei.bie@intel.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
5 years agovhost/test: fix build for vhost test
Tiwei Bie [Wed, 28 Aug 2019 05:36:59 +0000 (13:36 +0800)]
vhost/test: fix build for vhost test

Since below commit, callers need to specify the iov_limit in
vhost_dev_init() explicitly.

Fixes: b46a0bf78ad7 ("vhost: fix OOB in get_rx_bufs()")
Cc: stable@vger.kernel.org
Signed-off-by: Tiwei Bie <tiwei.bie@intel.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
5 years agoALSA: hda/realtek - Fix the problem of two front mics on a ThinkCentre
Hui Wang [Wed, 4 Sep 2019 05:53:27 +0000 (13:53 +0800)]
ALSA: hda/realtek - Fix the problem of two front mics on a ThinkCentre

This ThinkCentre machine has a new realtek codec alc222, it is not
in the support list, we add it in the realtek.c then this machine
can apply FIXUPs for the realtek codec.

And this machine has two front mics which can't be handled
by PA so far, it uses the pin 0x18 and 0x19 as the front mics, as
a result the existing FIXUP ALC294_FIXUP_LENOVO_MIC_LOCATION doesn't
work on this machine. Fortunately another FIXUP
ALC283_FIXUP_HEADSET_MIC also can change the location for one of the
two mics on this machine.

Link: https://lore.kernel.org/r/20190904055327.9883-1-hui.wang@canonical.com
Signed-off-by: Hui Wang <hui.wang@canonical.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
5 years agodmaengine: rcar-dmac: Fix DMACHCLR handling if iommu is mapped
Yoshihiro Shimoda [Mon, 2 Sep 2019 11:44:03 +0000 (20:44 +0900)]
dmaengine: rcar-dmac: Fix DMACHCLR handling if iommu is mapped

The commit 20c169aceb45 ("dmaengine: rcar-dmac: clear pertinence
number of channels") forgets to clear the last channel by
DMACHCLR in rcar_dmac_init() (and doesn't need to clear the first
channel) if iommu is mapped to the device. So, this patch fixes it
by using "channels_mask" bitfield.

Note that the hardware and driver don't support more than 32 bits
in DMACHCLR register anyway, so this patch should reject more than
32 channels in rcar_dmac_parse_of().

Fixes: 20c169aceb459575 ("dmaengine: rcar-dmac: clear pertinence number of channels")
Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Reviewed-by: Simon Horman <horms+renesas@verge.net.au>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/1567424643-26629-1-git-send-email-yoshihiro.shimoda.uh@renesas.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
5 years agodmaengine: sprd: Fix the DMA link-list configuration
Baolin Wang [Fri, 30 Aug 2019 07:37:45 +0000 (15:37 +0800)]
dmaengine: sprd: Fix the DMA link-list configuration

For the Spreadtrum DMA link-list mode, when the DMA engine got a slave
hardware request, which will trigger the DMA engine to load the DMA
configuration from the link-list memory automatically. But before the
slave hardware request, the slave will get an incorrect residue due
to the first node used to trigger the link-list was configured as the
last source address and destination address.

Thus we should make sure the first node was configured the start source
address and destination address, which can fix this issue.

Fixes: 4ac695464763 ("dmaengine: sprd: Support DMA link-list mode")
Signed-off-by: Baolin Wang <baolin.wang@linaro.org>
Link: https://lore.kernel.org/r/77868edb7aff9d5cb12ac3af8827ef2e244441a6.1567150471.git.baolin.wang@linaro.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
5 years agoASoC: es8316: add clock control of MCLK
Katsuhiro Suzuki [Tue, 3 Sep 2019 16:53:20 +0000 (01:53 +0900)]
ASoC: es8316: add clock control of MCLK

This patch introduce clock property for MCLK master freq control.
Driver will set rate of MCLK master if set_sysclk is called and
changing sysclk by board driver.

[Modified slightly to apply without an earlier patch in the series due
to context diffs -- broonie]

Signed-off-by: Katsuhiro Suzuki <katsuhiro@katsuster.net>
Link: https://lore.kernel.org/r/20190903165322.20791-2-katsuhiro@katsuster.net
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: es8316: add DT-bindings
Katsuhiro Suzuki [Tue, 3 Sep 2019 16:53:22 +0000 (01:53 +0900)]
ASoC: es8316: add DT-bindings

This patch adds missing DT-bindings document for Everest ES8316.

Signed-off-by: Katsuhiro Suzuki <katsuhiro@katsuster.net>
Link: https://lore.kernel.org/r/20190903165322.20791-4-katsuhiro@katsuster.net
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoRevert "mmc: core: do not retry CMD6 in __mmc_switch()"
Jan Kaisrlik [Tue, 20 Aug 2019 11:42:29 +0000 (13:42 +0200)]
Revert "mmc: core: do not retry CMD6 in __mmc_switch()"

Turns out the commit 3a0681c7448b ("mmc: core: do not retry CMD6 in
__mmc_switch()") breaks initialization of a Toshiba THGBMNG5 eMMC card,
when using the meson-gx-mmc.c driver on a custom board based on Amlogic
A113D.

The CMD6 that switches the card into HS200 mode is then one that fails and
according to the below printed messages from the log:

[    1.648951] mmc0: mmc_select_hs200 failed, error -84
[    1.648988] mmc0: error -84 whilst initialising MMC card

After some analyze, it turns out that adding a delay of ~5ms inside
mmc_select_bus_width() but after mmc_compare_ext_csds() has been executed,
also fixes the problem. Adding yet some more debug code, trying to figure
out if potentially the card could be in a busy state, both by using CMD13
and ->card_busy() ops concluded that this was not the case.

Therefore, let's simply revert the commit that dropped support for retrying
of CMD6, as this also fixes the problem.

Fixes: 3a0681c7448b ("mmc: core: do not retry CMD6 in __mmc_switch()")
Cc: stable@vger.kernel.org
Signed-off-by: Jan Kaisrlik <ja.kaisrlik@gmail.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
5 years agoiommu/vt-d: Remove global page flush support
Jacob Pan [Mon, 26 Aug 2019 15:53:29 +0000 (08:53 -0700)]
iommu/vt-d: Remove global page flush support

Global pages support is removed from VT-d spec 3.0. Since global pages G
flag only affects first-level paging structures and because DMA request
with PASID are only supported by VT-d spec. 3.0 and onward, we can
safely remove global pages support.

For kernel shared virtual address IOTLB invalidation, PASID
granularity and page selective within PASID will be used. There is
no global granularity supported. Without this fix, IOTLB invalidation
will cause invalid descriptor error in the queued invalidation (QI)
interface.

Fixes: 1c4f88b7f1f9 ("iommu/vt-d: Shared virtual address in scalable mode")
Reported-by: Sanjay K Kumar <sanjay.k.kumar@intel.com>
Signed-off-by: Jacob Pan <jacob.jun.pan@linux.intel.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
5 years agoASoC: audio-graph: indicate rebind issue
Kuninori Morimoto [Tue, 3 Sep 2019 08:16:05 +0000 (17:16 +0900)]
ASoC: audio-graph: indicate rebind issue

ALSA SoC try to rebind Sound Card if Card/CPU/Codec/Platform
were unbinded and re-binded again.
But, Audio Graph Card might can't rebind again if user do for example

unbind CPU or Codec driver
bind   CPU or Codec driver

Because Audio Graph Card is still pointing old/unbinded
CPU or Codec driver's DAI name at dlc->dai_name.

To avoid this issue, it needs to alloc memory and keep DAI name
even though if CPU or Codec driver was unbinded.
Or, always do unbind/bind at Sound Card.

For now, this patch indicates this issue as FIXME.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/87sgpdu75m.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoASoC: simple-card: indicate rebind issue
Kuninori Morimoto [Tue, 3 Sep 2019 08:15:54 +0000 (17:15 +0900)]
ASoC: simple-card: indicate rebind issue

ALSA SoC try to rebind Sound Card if Card/CPU/Codec/Platform
were unbinded and re-binded again.
But, Simple Card might can't rebind again if user do for example

unbind CPU or Codec driver
bind   CPU or Codec driver

Because Simple Card is still pointing old/unbinded
CPU or Codec driver's DAI name at dlc->dai_name.

To avoid this issue, it needs to alloc memory and keep DAI name
even though if CPU or Codec driver was unbinded.
Or, always do unbind/bind at Sound Card.

For now, this patch indicates this issue as FIXME.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/87tv9tu75x.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoMerge branch 'linux-5.3' of git://github.com/skeggsb/linux into drm-fixes
Dave Airlie [Tue, 3 Sep 2019 07:03:35 +0000 (17:03 +1000)]
Merge branch 'linux-5.3' of git://github.com/skeggsb/linux into drm-fixes

Single nouveau firmware fix.

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Ben Skeggs <skeggsb@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/CACAvsv5uGLgDY8V8pWgEH0-YhkCEgvHE=NZ1W_m0gJaoFPuQ0g@mail.gmail.com