October 2015 - Alsa-devel - mailman.alsa-project.org

Re: [alsa-devel] [PATCH] ALSA: seq_oss: fix waitqueue_active without memory barrier in snd-seq-oss
by Takashi Iwai 09 Oct '15

09 Oct '15

On Fri, 09 Oct 2015 02:35:54 +0200, Kosuke Tatsukawa wrote: > > snd_seq_oss_readq_put_event() seems to be missing a memory barrier which > might cause the waker to not notice the waiter and miss sending a > wake_up as in the following figure. > > snd_seq_oss_readq_put_event snd_seq_oss_readq_wait > ------------------------------------------------------------------------ > /* wait_event_interruptible_timeout */ > /* __wait_event_interruptible_timeout */ > /* ___wait_event */ > for (;;) { prepare_to_wait_event(&wq, &__wait, > state); > spin_lock_irqsave(&q->lock, flags); > if (waitqueue_active(&q->midi_sleep)) > /* The CPU might reorder the test for > the waitqueue up here, before > prior writes complete */ > if ((q->qlen>0 || q->head==q->tail) > ... > __ret = schedule_timeout(__ret) > if (q->qlen >= q->maxlen - 1) { > memcpy(&q->q[q->tail], ev, sizeof(*ev)); > q->tail = (q->tail + 1) % q->maxlen; > q->qlen++; > ------------------------------------------------------------------------ > > There are two other place in sound/core/seq/oss/ which have similar > code. The attached patch removes the call to waitqueue_active() leaving > just wake_up() behind. This fixes the problem because the call to > spin_lock_irqsave() in wake_up() will be an ACQUIRE operation. > > I found this issue when I was looking through the linux source code > for places calling waitqueue_active() before wake_up*(), but without > preceding memory barriers, after sending a patch to fix a similar > issue in drivers/tty/n_tty.c (Details about the original issue can be > found here: https://lkml.org/lkml/2015/9/28/849). > > Signed-off-by: Kosuke Tatsukawa <tatsu(a)ab.jp.nec.com> Thanks, applied now. Takashi > --- > sound/core/seq/oss/seq_oss_readq.c | 6 ++---- > sound/core/seq/oss/seq_oss_writeq.c | 4 +--- > 2 files changed, 3 insertions(+), 7 deletions(-) > > diff --git a/sound/core/seq/oss/seq_oss_readq.c b/sound/core/seq/oss/seq_oss_readq.c > index ccd8935..046cb58 100644 > --- a/sound/core/seq/oss/seq_oss_readq.c > +++ b/sound/core/seq/oss/seq_oss_readq.c > @@ -91,8 +91,7 @@ snd_seq_oss_readq_clear(struct seq_oss_readq *q) > q->head = q->tail = 0; > } > /* if someone sleeping, wake'em up */ > - if (waitqueue_active(&q->midi_sleep)) > - wake_up(&q->midi_sleep); > + wake_up(&q->midi_sleep); > q->input_time = (unsigned long)-1; > } > > @@ -138,8 +137,7 @@ snd_seq_oss_readq_put_event(struct seq_oss_readq *q, union evrec *ev) > q->qlen++; > > /* wake up sleeper */ > - if (waitqueue_active(&q->midi_sleep)) > - wake_up(&q->midi_sleep); > + wake_up(&q->midi_sleep); > > spin_unlock_irqrestore(&q->lock, flags); > > diff --git a/sound/core/seq/oss/seq_oss_writeq.c b/sound/core/seq/oss/seq_oss_writeq.c > index d50338b..1f6788a 100644 > --- a/sound/core/seq/oss/seq_oss_writeq.c > +++ b/sound/core/seq/oss/seq_oss_writeq.c > @@ -138,9 +138,7 @@ snd_seq_oss_writeq_wakeup(struct seq_oss_writeq *q, abstime_t time) > spin_lock_irqsave(&q->sync_lock, flags); > q->sync_time = time; > q->sync_event_put = 0; > - if (waitqueue_active(&q->sync_sleep)) { > - wake_up(&q->sync_sleep); > - } > + wake_up(&q->sync_sleep); > spin_unlock_irqrestore(&q->sync_lock, flags); > } > > -- > 1.7.1 >

1 0

[alsa-devel] Release of the tool for benchmarking sound server power efficiency
by Alexander E. Patrakov 09 Oct '15

09 Oct '15

Earlier today (sorry - yesterday!) at the audio mini-summit I promised to release a tool that I used for evaluating sound server energy efficiency. The tool is a shell script, which is attached, together with an example report. It needs to be run as root (in a throwaway VM, if you need), with one argument: the name of a directory to be created and populated with various files. The script needs the "debootstrap" program to be installed. The script will populate the directory with a Debian Jessie amd64 installation, compile CRAS there, and pack everything into an initramfs file that can be used for evaluation. It will produce two files as the main result: vmlinuz and initramfs.img, near the directory that it creates. The initramfs will play a sound file in a loop (please use headphones, so that your laptop does not annoy anyone) and log the laptop battery energy or charge. If it drops below 75% of the initial value, the experiment is terminated, the log is saved to the disk, and the laptop is powered off. You can add this to your boot loader. Example for extlinux: LABEL battery-pulse7 MENU LABEL Battery Benchmark (pulseaudio 7) LINUX /boot/battery-benchmark/vmlinuz INITRD /boot/battery-benchmark/initramfs.img APPEND quiet root=ram save=/dev/md126p1 test_case=pulse pa_version=7 latency_msec=28 dobenchmark Here is the meaning of the arguments: quiet: don't print verbose messages to the console root=ram: use ramfs (not tmpfs) for the extracted initramfs save=/dev/...: save the log to this block device at the end. The /battery-reports directory will be created, and the log will be placed there. FAT32, NTFS and EXT4 filesystems are supported. test_case=pulse|alsa|cras: what we want to test. alsa means raw hw device. pa_version=5|6|7: version of PulseAudio to test latency_msec: buffer size or its equivalent, in milliseconds cras_output: the correct output node for CRAS. Usually 5:0 (the default) or 6:0, please see cras_test_client --dump_server_info resampler: force a given resampler for PulseAudio (if not given, no resampler is used) dobenchmark: dpms-blank the screen and perform the measurement automatically To replicate the results in my presentation, it is sufficient to obtain logs for all three versions of PulseAudio, as well as ALSA and CRAS, for the following latency_msec values: 7, 28, 448. Please send any reports created by this tool to me directly, and mention the laptop model. A big "thanks" goes to Felipe Sateler for providing Debian packages of PulseAudio that are mostly compatible with Debian Jessie. -- Alexander E. Patrakov

1 0

[alsa-devel] [PATCH alsa-utils] build: Do not try to detect cross-compiler
by Mike Frysinger 08 Oct '15

08 Oct '15

From: Mike Frysinger <vapier(a)chromium.org> cross compilers are passed via path may not be a gcc based cross compiler in such cases this check fails and try's to force gcc based cross compiler detection, This code is a convenience that limits the build system. Signed-off-by: Mike Frysinger <vapier(a)gentoo.org> --- Note: this is the same fix that landed in alsa-lib. configure.ac | 14 -------------- 1 file changed, 14 deletions(-) diff --git a/configure.ac b/configure.ac index 8c2d1a5..79a9aaf 100644 --- a/configure.ac +++ b/configure.ac @@ -12,20 +12,6 @@ AM_GNU_GETTEXT_VERSION([0.15]) dnl Checks for programs. -dnl try to gues cross-compiler if not set -if test "x$target" != "x$host" -a -z "`echo $CC | grep -e '-gcc'`"; -then - AC_MSG_CHECKING(for cross-compiler) - - which ${program_prefix}gcc >/dev/null 2>&1 && CC=${program_prefix}gcc - which ${target_cpu}-${target_os}-gcc >/dev/null 2>&1 \ - && CC=${target_cpu}-${target-os}-gcc - which ${target_cpu}-${target_vendor}-${target_os}-gcc >/dev/null 2>&1 \ - && CC=${target_cpu}-${target_vendor}-${target_os}-gcc - - AC_MSG_RESULT($CC) -fi - AC_PROG_CC dnl AC_PROG_CXX AC_PROG_INSTALL -- 2.5.2

2 1

[alsa-devel] [PATCH 1/2] ALSA: hdac: Copy codec helpers to core
by Vinod Koul 08 Oct '15

08 Oct '15

From: "Subhransu S. Prusty" <subhransu.s.prusty(a)intel.com> The current codec helpers are local to hda code and needs to be moved to core so that other users can use it. The helpers to read/write the codec and to check the power state of widgets is copied Signed-off-by: Subhransu S. Prusty <subhransu.s.prusty(a)intel.com> Signed-off-by: Vinod Koul <vinod.koul(a)intel.com> --- include/sound/hdaudio.h | 6 ++++ sound/hda/hdac_device.c | 81 +++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 87 insertions(+) diff --git a/include/sound/hdaudio.h b/include/sound/hdaudio.h index 49bc836fcd84..26e956f4b7c6 100644 --- a/include/sound/hdaudio.h +++ b/include/sound/hdaudio.h @@ -147,6 +147,12 @@ int snd_hdac_query_supported_pcm(struct hdac_device *codec, hda_nid_t nid, bool snd_hdac_is_supported_format(struct hdac_device *codec, hda_nid_t nid, unsigned int format); +int snd_hdac_codec_read(struct hdac_device *hdac, hda_nid_t nid, + int flags, unsigned int verb, unsigned int parm); +int snd_hdac_codec_write(struct hdac_device *hdac, hda_nid_t nid, + int flags, unsigned int verb, unsigned int parm); +bool snd_hdac_check_power_state(struct hdac_device *hdac, + hda_nid_t nid, unsigned int target_state); /** * snd_hdac_read_parm - read a codec parameter * @codec: the codec object diff --git a/sound/hda/hdac_device.c b/sound/hda/hdac_device.c index db96042a497f..b3b0ad289df1 100644 --- a/sound/hda/hdac_device.c +++ b/sound/hda/hdac_device.c @@ -952,3 +952,84 @@ bool snd_hdac_is_supported_format(struct hdac_device *codec, hda_nid_t nid, return true; } EXPORT_SYMBOL_GPL(snd_hdac_is_supported_format); + +static unsigned int codec_read(struct hdac_device *hdac, hda_nid_t nid, + int flags, unsigned int verb, unsigned int parm) +{ + unsigned int cmd = snd_hdac_make_cmd(hdac, nid, verb, parm); + unsigned int res; + + if (snd_hdac_exec_verb(hdac, cmd, flags, &res)) + return -1; + + return res; +} + +static int codec_write(struct hdac_device *hdac, hda_nid_t nid, + int flags, unsigned int verb, unsigned int parm) +{ + unsigned int cmd = snd_hdac_make_cmd(hdac, nid, verb, parm); + + return snd_hdac_exec_verb(hdac, cmd, flags, NULL); +} + +/** + * snd_hdac_codec_read - send a command and get the response + * @hdac: the HDAC device + * @nid: NID to send the command + * @flags: optional bit flags + * @verb: the verb to send + * @parm: the parameter for the verb + * + * Send a single command and read the corresponding response. + * + * Returns the obtained response value, or -1 for an error. + */ +int snd_hdac_codec_read(struct hdac_device *hdac, hda_nid_t nid, + int flags, unsigned int verb, unsigned int parm) +{ + return codec_read(hdac, nid, flags, verb, parm); +} +EXPORT_SYMBOL_GPL(snd_hdac_codec_read); + +/** + * snd_hdac_codec_write - send a single command without waiting for response + * @hdac: the HDAC device + * @nid: NID to send the command + * @flags: optional bit flags + * @verb: the verb to send + * @parm: the parameter for the verb + * + * Send a single command without waiting for response. + * + * Returns 0 if successful, or a negative error code. + */ +int snd_hdac_codec_write(struct hdac_device *hdac, hda_nid_t nid, + int flags, unsigned int verb, unsigned int parm) +{ + return codec_write(hdac, nid, flags, verb, parm); +} +EXPORT_SYMBOL_GPL(snd_hdac_codec_write); + +/* + * snd_hdac_check_power_state: check whether the actual power state matches + * with the target state + * + * @hdac: the HDAC device + * @nid: NID to send the command + * @target_state: target state to check for + * + * Return true if state matches, false if not + */ +bool snd_hdac_check_power_state(struct hdac_device *hdac, + hda_nid_t nid, unsigned int target_state) +{ + unsigned int state = codec_read(hdac, nid, 0, + AC_VERB_GET_POWER_STATE, 0); + + if (state & AC_PWRST_ERROR) + return true; + state = (state >> 4) & 0x0f; + return (state == target_state); +} +EXPORT_SYMBOL_GPL(snd_hdac_check_power_state); -- 2.4.3

2 2

[alsa-devel] [PATCH] ASoC: Document DAI signal polarity
by Anatol Pomozov 08 Oct '15

08 Oct '15

Currently there is no clear definition of what FSYNC polarity is. Different drivers use its own definition of what is "normal" and what is "inverted" fsync. This leads to compatibility problems between drivers. For example TegraX1 driver assumes that DSP-A format with frames starting at rising FSYNC edge has "inverted" polarity, while RT5677 assumes it is "normal" polarity. Explicitly specify meaning of BCLK/FSYNC polarity to avoid future compatibility problems. Signed-off-by: Anatol Pomozov <anatol.pomozov(a)gmail.com> --- include/sound/soc-dai.h | 17 ++++++++++++++++- 1 file changed, 16 insertions(+), 1 deletion(-) diff --git a/include/sound/soc-dai.h b/include/sound/soc-dai.h index 2df96b1..91e2e61 100644 --- a/include/sound/soc-dai.h +++ b/include/sound/soc-dai.h @@ -48,10 +48,25 @@ struct snd_compr_stream; #define SND_SOC_DAIFMT_GATED (0 << 4) /* clock is gated */ /* - * DAI hardware signal inversions. + * DAI hardware signal polarity. * * Specifies whether the DAI can also support inverted clocks for the specified * format. + * + * BCLK: + * - "normal" polarity means signal is available at rising edge of BCLK + * - "inverted" polarity means signal is available at falling edge of BCLK + * + * FSYNC "normal" polarity depends on the frame format: + * - I2S: frame consists of left then right channel data. Left channel starts + * with falling FSYNC edge, right channel starts with rising FSYNC edge. + * - Left/Right Justified: frame consists of left then right channel data. + * Left channel starts with rising FSYNC edge, right channel starts with + * falling FSYNC edge. + * - DSP A/B: Frame starts with rising FSYNC edge. + * - AC97: Frame starts with rising FSYNC edge. + * + * "Negative" FSYNC polarity is the one opposite of "normal" polarity. */ #define SND_SOC_DAIFMT_NB_NF (0 << 8) /* normal bit clock + frame */ #define SND_SOC_DAIFMT_NB_IF (2 << 8) /* normal BCLK + inv FRM */ -- 2.6.0.rc2.230.g3dd15c0

1 0

[alsa-devel] [PATCH v2 0/3] ASoC: da7213: Add DT support for codec
by Adam Thomson 08 Oct '15

08 Oct '15

This patch set adds DT support to the codec driver, and supporting bindings documentation. Patch set is based against v4.3-rc4 Changes in v2: - Split mclk data support changes into separate patch - Use of_match_ptr() for match table assignment - Only call clk_prepare_enable/disable() if mclk data provided (not NULL). - Add error checking for clk_prepare_enable() usage. Adam Thomson (3): ASoC: da7213: Add DT support to codec driver ASoC: da7213: Add support to handle mclk data provided to driver ASoC: da7213: Add bindings documentation for codec driver Documentation/devicetree/bindings/sound/da7213.txt | 41 +++++ include/sound/da7213.h | 3 - sound/soc/codecs/da7213.c | 190 +++++++++++++++++++-- sound/soc/codecs/da7213.h | 8 +- 4 files changed, 221 insertions(+), 21 deletions(-) create mode 100644 Documentation/devicetree/bindings/sound/da7213.txt -- 1.9.3

5 9

[alsa-devel] USB audio device startup question
by Ricard Wanderlof 08 Oct '15

08 Oct '15

(This is in reference to getting the Zoom R16 to work, but the main query in this post is of a more general nature). In trying to get a Zoom R16 in audio interface mode to work in Linux I'm trying to understand how the USB-audio (snd-usb-audio module) subsystem works in Linux. I've looked in the USB audio version 1 and 2 documents, as well as the USB 2.0 standard, but there are some things that don't seem to be well defined there, like exactly how the audio data is packaged in the isochronous transfer packets. I've been using Wireshark to visualize the USB communication, and also comparing with the corresponding transfers on a Windows box by dumping the traffic with USBPcap. The only other USB audio interface I have access to here is an Edirol UA-5. I've noted that when plugging in the UA-5, there's period of about 15 seconds during which capture and playback are set up and audio data transferred, after which the interface is shut down again. When enabling the R16, during the corresponding initial period I see playback data but no capture data during this phase. I don't know if that is significant or not, for instance, if the lack of capture data during this phase is because the playback is not working as it should. My question is: what governs this startup phase, can someone point to a specific part of the code or some documentation describing what is supposed to go on here? Is it some sort of device test period to determine that the device actually works? The problem with the R16 is that although capture works fine, playback doesn't result in any audio. With the correct quirk set up (most importantly, using a 24 bit format), aplay appears to work fine with no errors, neither are there any complaints from dmesg, and looking at the communication I can see outgoing isochronous data from the host being transferred fine. More work needs to be done, but I've looked at the startup of the Windows driver which is deceptively simple: it basically just sets the appropriate capture and playback interfaces, together with sending off some sample rate USB control commands, the isochronous link comes up and that seems to be it. I've tried to emulate the specific USB control commands in Linux, as well as add some delays that there appear to be in the Windows driver, but so far it hasn't made a difference. /Ricard -- Ricard Wolf Wanderlöf ricardw(at)axis.com Axis Communications AB, Lund, Sweden www.axis.com Phone +46 46 272 2016 Fax +46 46 13 61 30

2 1

[alsa-devel] [PATCH] ASoC: rt5645: Remove __initdata annotation from dmi_platform_intel_broadwell
by Jarkko Nikula 08 Oct '15

08 Oct '15

Fix the "Section mismatch" warning when this driver is built as module by removing the __initdata annotation from dmi_platform_intel_broadwell[]. Signed-off-by: Jarkko Nikula <jarkko.nikula(a)linux.intel.com> --- sound/soc/codecs/rt5645.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/soc/codecs/rt5645.c b/sound/soc/codecs/rt5645.c index 4c4fe6b456d5..89ba4ffdbee7 100644 --- a/sound/soc/codecs/rt5645.c +++ b/sound/soc/codecs/rt5645.c @@ -3244,7 +3244,7 @@ static int buddy_quirk_cb(const struct dmi_system_id *id) return 1; } -static struct dmi_system_id dmi_platform_intel_broadwell[] __initdata = { +static struct dmi_system_id dmi_platform_intel_broadwell[] = { { .ident = "Chrome Buddy", .callback = buddy_quirk_cb, -- 2.6.1

1 1

[alsa-devel] [tinycompress] [PATCH] build: Fix make distcheck
by arun＠accosted.net 08 Oct '15

08 Oct '15

From: Arun Raghavan <git(a)arunraghavan.net> srcdir may not be the same as builddir, so use the correct variable for the include path. --- src/lib/Makefile.am | 2 +- src/utils/Makefile.am | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/src/lib/Makefile.am b/src/lib/Makefile.am index 6f900ac..4baeab5 100644 --- a/src/lib/Makefile.am +++ b/src/lib/Makefile.am @@ -2,4 +2,4 @@ tinycompressdir = $(libdir) tinycompress_LTLIBRARIES = libtinycompress.la libtinycompress_la_SOURCES = compress.c -libtinycompress_la_CFLAGS = -I $(top_builddir)/include +libtinycompress_la_CFLAGS = -I$(top_srcdir)/include diff --git a/src/utils/Makefile.am b/src/utils/Makefile.am index 6a89c45..5f685e3 100644 --- a/src/utils/Makefile.am +++ b/src/utils/Makefile.am @@ -3,8 +3,8 @@ bin_PROGRAMS = cplay crecord cplay_SOURCES = cplay.c crecord_SOURCES = crec.c -cplay_CFLAGS = -I$(top_builddir)/include -crecord_CFLAGS = -I$(top_builddir)/include +cplay_CFLAGS = -I$(top_srcdir)/include +crecord_CFLAGS = -I$(top_srcdir)/include cplay_LDADD = $(top_builddir)/src/lib/libtinycompress.la -- 2.4.3

2 1

[alsa-devel] [PATCH 0/2] ASoC: rockchip: i2s: add 8 channels capture and lrck-mode
by Sugar Zhang 08 Oct '15

08 Oct '15

Support max 8 channels capture. support lrck clk mode configuration, there are 3 modes: - txrx: lrck_tx and lrck_rx are different. - tx_share: lrck_tx is shared with lrck_rx. - rx_share: lrck_rx is shared with lrck_tx. Sugar Zhang (2): ASoC: rockchip: i2s: add 8 channels capture and lrck-mode support Documentation: DT bindings: rockchip-i2s: add capture and lrck-mode .../devicetree/bindings/sound/rockchip-i2s.txt | 5 +++ sound/soc/rockchip/rockchip_i2s.c | 48 +++++++++++++++++++++- sound/soc/rockchip/rockchip_i2s.h | 23 +++++++++++ 3 files changed, 74 insertions(+), 2 deletions(-) -- 2.3.6

4 16