December 2016 - Alsa-devel - mailman.alsa-project.org

[alsa-devel] [PATCH v4] ASoC: omap-mcbsp: Add PM QoS support for McBSP to prevent glitches
by Matt Ranostay 03 Jan '17

03 Jan '17

We can get audio errors if hitting deeper idle states on omaps: [alsa.c:230] error: Fatal problem with alsa output, error -5. [audio.c:614] error: Error in writing audio (Input/output error?)! This seems to happen with off mode idle enabled as power for the whole SoC may get cut off between filling the McBSP fifo using DMA. While active DMA blocks deeper idle states in hardware, McBSP activity does not seem to do so. Basing the QoS latency calculation on the FIFO size, threshold, sample rate, and channels. Based on the original patch by Tony Lindgren Link: https://patchwork.kernel.org/patch/9305867/ Cc: Tony Lindgren <tony(a)atomide.com> Cc: Peter Ujfalusi <peter.ujfalusi(a)ti.com> Signed-off-by: Matt Ranostay <matt(a)ranostay.consulting> --- Changes from v1: * add calculations for latency per number of FIFO locations Changes from v2: * add missing mcbsp.h header change Changes from v3: * base the latency calculations on threshold, buffer size, sample rate, and channels sound/soc/omap/mcbsp.c | 19 +++++++++++++++++++ sound/soc/omap/mcbsp.h | 3 +++ sound/soc/omap/omap-mcbsp.c | 21 ++++++++++++++++++++- 3 files changed, 42 insertions(+), 1 deletion(-) diff --git a/sound/soc/omap/mcbsp.c b/sound/soc/omap/mcbsp.c index 06fec5699cc8..8f792e714894 100644 --- a/sound/soc/omap/mcbsp.c +++ b/sound/soc/omap/mcbsp.c @@ -25,6 +25,7 @@ #include <linux/io.h> #include <linux/slab.h> #include <linux/pm_runtime.h> +#include <linux/pm_qos.h> #include <linux/platform_data/asoc-ti-mcbsp.h> @@ -640,9 +641,19 @@ void omap_mcbsp_free(struct omap_mcbsp *mcbsp) */ void omap_mcbsp_start(struct omap_mcbsp *mcbsp, int tx, int rx) { + struct pm_qos_request *pm_qos_req = &mcbsp->pm_qos_req; int enable_srg = 0; u16 w; + /* Prevent omap hardware from hitting off between fifo fills */ + if (mcbsp->latency) { + if (pm_qos_request_active(pm_qos_req)) + pm_qos_update_request(pm_qos_req, mcbsp->latency); + else + pm_qos_add_request(pm_qos_req, PM_QOS_CPU_DMA_LATENCY, + mcbsp->latency); + } + if (mcbsp->st_data) omap_st_start(mcbsp); @@ -731,6 +742,11 @@ void omap_mcbsp_stop(struct omap_mcbsp *mcbsp, int tx, int rx) if (mcbsp->st_data) omap_st_stop(mcbsp); + + if (mcbsp->active == 1 && pm_qos_request_active(&mcbsp->pm_qos_req)) { + pm_qos_remove_request(&mcbsp->pm_qos_req); + mcbsp->latency = 0; + } } int omap2_mcbsp_set_clks_src(struct omap_mcbsp *mcbsp, u8 fck_src_id) @@ -1098,6 +1114,9 @@ int omap_mcbsp_init(struct platform_device *pdev) void omap_mcbsp_cleanup(struct omap_mcbsp *mcbsp) { + if (pm_qos_request_active(&mcbsp->pm_qos_req)) + pm_qos_remove_request(&mcbsp->pm_qos_req); + if (mcbsp->pdata->buffer_size) sysfs_remove_group(&mcbsp->dev->kobj, &additional_attr_group); diff --git a/sound/soc/omap/mcbsp.h b/sound/soc/omap/mcbsp.h index 61e93b1c185d..d312e72f7979 100644 --- a/sound/soc/omap/mcbsp.h +++ b/sound/soc/omap/mcbsp.h @@ -323,8 +323,11 @@ struct omap_mcbsp { unsigned int fmt; unsigned int in_freq; + unsigned int latency; int clk_div; int wlen; + + struct pm_qos_request pm_qos_req; }; void omap_mcbsp_config(struct omap_mcbsp *mcbsp, diff --git a/sound/soc/omap/omap-mcbsp.c b/sound/soc/omap/omap-mcbsp.c index d018e966e533..ddaec37fb4c3 100644 --- a/sound/soc/omap/omap-mcbsp.c +++ b/sound/soc/omap/omap-mcbsp.c @@ -226,6 +226,7 @@ static int omap_mcbsp_dai_hw_params(struct snd_pcm_substream *substream, int wlen, channels, wpf; int pkt_size = 0; unsigned int format, div, framesize, master; + unsigned int buffer_size = mcbsp->pdata->buffer_size; dma_data = snd_soc_dai_get_dma_data(cpu_dai, substream); channels = params_channels(params); @@ -240,7 +241,9 @@ static int omap_mcbsp_dai_hw_params(struct snd_pcm_substream *substream, default: return -EINVAL; } - if (mcbsp->pdata->buffer_size) { + if (buffer_size) { + int latency; + if (mcbsp->dma_op_mode == MCBSP_DMA_MODE_THRESHOLD) { int period_words, max_thrsh; int divider = 0; @@ -271,6 +274,22 @@ static int omap_mcbsp_dai_hw_params(struct snd_pcm_substream *substream, /* Use packet mode for non mono streams */ pkt_size = channels; } + + latency = ((((buffer_size - pkt_size) / channels) * 1000) + / (params->rate_num / params->rate_den)); + + /* + * Use the lowest QoS latency value for active streams, but + * this has the disadvantage of using lower than needed latency + * if an stream with lower QoS ends before one with a higher + * one. Latency setting only gets reset when all streams + * complete. + */ + if (latency <= mcbsp->latency) + mcbsp->latency = latency; + else + mcbsp->latency = 0; + omap_mcbsp_set_threshold(substream, pkt_size); } -- 2.7.4

2 2

[alsa-devel] [PATCH][alsa-lib] topology: fix unused-const-variable warning
by Takashi Sakamoto 02 Jan '17

02 Jan '17

Last year, unused static const variable was added, then compiler generates a below warning. dapm.c:43:30: warning: ‘widget_control_map’ defined but not used [-Wunused-const-variable=] static const struct map_elem widget_control_map[] = { ^~~~~~~~~~~~~~~~~~ This commit removes it. Fixes: 01a0e1a1c219 ("topology: Add DAPM object parser") Signed-off-by: Takashi Sakamoto <o-takashi(a)sakamocchi.jp> --- src/topology/dapm.c | 8 -------- 1 file changed, 8 deletions(-) diff --git a/src/topology/dapm.c b/src/topology/dapm.c index e830751..8c585a7 100644 --- a/src/topology/dapm.c +++ b/src/topology/dapm.c @@ -39,14 +39,6 @@ static const struct map_elem widget_map[] = { {"dai_link", SND_SOC_TPLG_DAPM_DAI_LINK}, }; -/* mapping of widget kcontrol text names to types */ -static const struct map_elem widget_control_map[] = { - {"volsw", SND_SOC_TPLG_DAPM_CTL_VOLSW}, - {"enum_double", SND_SOC_TPLG_DAPM_CTL_ENUM_DOUBLE}, - {"enum_virt", SND_SOC_TPLG_DAPM_CTL_ENUM_VIRT}, - {"enum_value", SND_SOC_TPLG_DAPM_CTL_ENUM_VALUE}, -}; - static int lookup_widget(const char *w) { unsigned int i; -- 2.9.3

4 8

[alsa-devel] [PATCH][alsa-lib] topology: fix sign-compare warning introduced to set_link_hw_config() and tplg_add_link_object()
by Takashi Sakamoto 02 Jan '17

02 Jan '17

Some functions recently introduced to this library include comparison between variables with different types, then compiler generates below warnings. pcm.c: In function ‘set_link_hw_config’: pcm.c:1179:16: warning: comparison between signed and unsigned integer expressions [-Wsign-compare] for (i = 0; i < cfg->tx_channels; i++) ^ pcm.c:1183:16: warning: comparison between signed and unsigned integer expressions [-Wsign-compare] for (i = 0; i < cfg->rx_channels; i++) ^ pcm.c: In function ‘tplg_add_link_object’: pcm.c:1223:16: warning: comparison between signed and unsigned integer expressions [-Wsign-compare] for (i = 0; i < link->num_streams; i++) ^ pcm.c:1231:16: warning: comparison between signed and unsigned integer expressions [-Wsign-compare] for (i = 0; i < link->num_hw_configs; i++) This commit suppresses them. Fixes: 6b4d775b9752 ("topology: Parse HW configurations of physical DAI links defined by C API") Signed-off-by: Takashi Sakamoto <o-takashi(a)sakamocchi.jp> --- src/topology/pcm.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/src/topology/pcm.c b/src/topology/pcm.c index 8f8a703..1de8655 100644 --- a/src/topology/pcm.c +++ b/src/topology/pcm.c @@ -1147,7 +1147,7 @@ int tplg_add_pcm_object(snd_tplg_t *tplg, snd_tplg_obj_template_t *t) static int set_link_hw_config(struct snd_soc_tplg_hw_config *cfg, struct snd_tplg_hw_config_template *tpl) { - int i; + unsigned int i; cfg->size = sizeof(*cfg); cfg->id = tpl->id; @@ -1190,7 +1190,7 @@ int tplg_add_link_object(snd_tplg_t *tplg, snd_tplg_obj_template_t *t) struct snd_tplg_link_template *link_tpl = t->link; struct snd_soc_tplg_link_config *link, *_link; struct tplg_elem *elem; - int i; + unsigned int i; if (t->type != SND_TPLG_TYPE_LINK && t->type != SND_TPLG_TYPE_BE && t->type != SND_TPLG_TYPE_CC) -- 2.9.3

3 3

[alsa-devel] [PATCH 0/2 v2] alsa-lib: dynamically adapt the avail_min on the slave
by Jiada Wang 02 Jan '17

02 Jan '17

When configuring avail_min to multiple of slave period size it can happen that user waits one slave period longer than needed for available data. Root cause is implicit grabbing of slave samples in avail_update operation. On next entering poll, the slave will wait for the avail_min threshold reached again, as he is not aware that there are already pending samples in the above layer which are not yet provided to user. Solution is to dynamically adapt the avail_min on the slave. Andreas Pape (2): plugin: dynamically update avail_min on slave rate: dynamic update avail_min on slave src/pcm/pcm_plugin.c | 64 +++++++++++++++++++++++++++++++++++++++++++++++++++- src/pcm/pcm_plugin.h | 3 +++ src/pcm/pcm_rate.c | 2 +- 3 files changed, 67 insertions(+), 2 deletions(-) -- 2.9.3

2 3

[alsa-devel] [PATCH 3/6] alsa-lib: pcm:dshare enable silence
by sutar.mounesh＠gmail.com 02 Jan '17

02 Jan '17

From: Alexander Jahn <ajahn(a)de.adit-jv.com> This issue depends on system load - if the process using dshare is scheduled fast enough, then there is no noise. A delay of e.g >~2ms produces hearable noise. Reproduction with instrumented aplay(sleep every 100th period for a given time): During the sleep time of 2000000us (2s) the hardware plays old samples in a loop before xrun is detected and recovered after the sleep. This is resolved by placing it in silence, in case of dshare plugin. Signed-off-by: Alexander Jahn <ajahn(a)de.adit-jv.com> Signed-off-by: Andreas Pape <apape(a)de.adit-jv.com> diff --git a/src/pcm/pcm_direct.c b/src/pcm/pcm_direct.c index 4234d66..779845c 100644 --- a/src/pcm/pcm_direct.c +++ b/src/pcm/pcm_direct.c @@ -1240,7 +1240,7 @@ int snd_pcm_direct_initialize_slave(snd_pcm_direct_t *dmix, snd_pcm_t *spcm, str return ret; } - if (dmix->type != SND_PCM_TYPE_DMIX) + if ((dmix->type != SND_PCM_TYPE_DMIX)&&(dmix->type != SND_PCM_TYPE_DSHARE)) goto __skip_silencing; ret = snd_pcm_sw_params_set_silence_threshold(spcm, &sw_params, 0); -- 1.7.9.5

2 1

[alsa-devel] [PATCH 2/6] alsa-lib: Fix for sync issue on xrun recover
by sutar.mounesh＠gmail.com 02 Jan '17

02 Jan '17

From: Andreas Pape <apape(a)de.adit-jv.com> If using very short periods, DSHARE/DSNOOP/DMIX may report underruns while in status 'prepared'. This prohibits correct recovery. Now slave xrun conditions for DSHARE/DSNOOP/DMIX are being handled properly. Signed-off-by: Andreas Pape <apape(a)de.adit-jv.com> Signed-off-by: Joshua Frkuska <joshua_frkuska(a)mentor.com> diff --git a/src/pcm/pcm_direct.c b/src/pcm/pcm_direct.c index 8f42b19..4234d66 100644 --- a/src/pcm/pcm_direct.c +++ b/src/pcm/pcm_direct.c @@ -572,6 +572,9 @@ int snd_pcm_direct_poll_revents(snd_pcm_t *pcm, struct pollfd *pfds, unsigned in } switch (snd_pcm_state(dmix->spcm)) { case SND_PCM_STATE_XRUN: + /*recover slave and update client state to xrun before returning POLLERR*/ + snd_pcm_direct_slave_recover(dmix); + snd_pcm_direct_client_chk_xrun(dmix, pcm); case SND_PCM_STATE_SUSPENDED: case SND_PCM_STATE_SETUP: events |= POLLERR; @@ -965,6 +968,83 @@ static void save_slave_setting(snd_pcm_direct_t *dmix, snd_pcm_t *spcm) #undef COPY_SLAVE +#define direct_sem_down_chk_ret(d, id) {\ + int semerr = snd_pcm_direct_semaphore_down(d, DIRECT_IPC_SEM_CLIENT); \ + if (semerr) {\ + SNDERR("SEMDOWN FAILED with err %d", semerr);\ + return semerr;\ + }\ +} + +#define direct_sem_up_chk_ret(d, id) {\ + int semerr = snd_pcm_direct_semaphore_up(d, DIRECT_IPC_SEM_CLIENT); \ + if (semerr) {\ + SNDERR("SEMUP FAILED with err %d", semerr);\ + return semerr;\ + }\ +} + +/* + * Recover slave on XRUN. + * Even if direct plugins disable xrun detection, there might be an xrun raised directly by some drivers. + * The first client recovers slave pcm. + * Each client needs to execute sw xrun handling afterwards + */ +int snd_pcm_direct_slave_recover(snd_pcm_direct_t *direct) +{ + int ret = 0; + + direct_sem_down_chk_ret(direct, DIRECT_IPC_SEM_CLIENT); + if (snd_pcm_state(direct->spcm) != SND_PCM_STATE_XRUN) { + /*ignore... someone else already did recovery*/ + direct_sem_up_chk_ret(direct, DIRECT_IPC_SEM_CLIENT); + return ret; + } + + ret = snd_pcm_prepare(direct->spcm); + if (ret < 0) { + SNDERR("recover: unable to prepare slave"); + direct_sem_up_chk_ret(direct, DIRECT_IPC_SEM_CLIENT); + return ret; + } + + if (direct->type == SND_PCM_TYPE_DSHARE) { + const snd_pcm_channel_area_t *dst_areas; + dst_areas = snd_pcm_mmap_areas(direct->spcm); + snd_pcm_areas_silence(dst_areas, 0, direct->spcm->channels, direct->spcm->buffer_size, direct->spcm->format); + } + + ret = snd_pcm_start(direct->spcm); + if (ret < 0) { + SNDERR("recover: unable to start slave"); + direct_sem_up_chk_ret(direct, DIRECT_IPC_SEM_CLIENT); + return ret; + } + direct->shmptr->recoveries++; + direct_sem_up_chk_ret(direct, DIRECT_IPC_SEM_CLIENT); + return 0; +} + +/* + * enter xrun state, if slave xrun occured + * @return: 0 - no xrun >0: xrun happened + */ +int snd_pcm_direct_client_chk_xrun(snd_pcm_direct_t *direct, snd_pcm_t *pcm) +{ + if (direct->shmptr->recoveries != direct->recoveries) { + /* no matter how many xruns we missed - so don't increment but just update to actual counter*/ + direct->recoveries = direct->shmptr->recoveries; + pcm->fast_ops->drop(pcm); + /*trigger_tstamp update is missing in drop callbacks*/ + gettimestamp(&direct->trigger_tstamp, pcm->tstamp_type); + /*no timer clear: if slave already entered xrun again the event is lost...*/ + /*snd_pcm_direct_clear_timer_queue(direct);*/ + direct->state = SND_PCM_STATE_XRUN; + return 1; + } + return 0; +} + /* * this function initializes hardware and starts playback operation with * no stop threshold (it operates all time without xrun checking) @@ -1380,6 +1460,7 @@ int snd_pcm_direct_open_secondary_client(snd_pcm_t **spcmp, snd_pcm_direct_t *dm dmix->slave_buffer_size = spcm->buffer_size; dmix->slave_period_size = dmix->shmptr->s.period_size; dmix->slave_boundary = spcm->boundary; + dmix->recoveries = dmix->shmptr->recoveries; ret = snd_pcm_mmap(spcm); if (ret < 0) { diff --git a/src/pcm/pcm_direct.h b/src/pcm/pcm_direct.h index 91e816c..569d2be 100644 --- a/src/pcm/pcm_direct.h +++ b/src/pcm/pcm_direct.h @@ -66,6 +66,7 @@ typedef struct { char socket_name[256]; /* name of communication socket */ snd_pcm_type_t type; /* PCM type (currently only hw) */ int use_server; + unsigned int recoveries; /* no of executed recoveries on slave*/ struct { unsigned int format; snd_interval_t rate; @@ -157,6 +158,7 @@ struct snd_pcm_direct { int var_periodsize; /* allow variable period size if max_periods is != -1*/ unsigned int channels; /* client's channels */ unsigned int *bindings; + unsigned int recoveries; /* mirror of executed recoveries on slave */ union { struct { int shmid_sum; /* IPC global sum ring buffer memory identification */ @@ -318,7 +320,8 @@ int snd_pcm_direct_open_secondary_client(snd_pcm_t **spcmp, snd_pcm_direct_t *dm snd_pcm_chmap_query_t **snd_pcm_direct_query_chmaps(snd_pcm_t *pcm); snd_pcm_chmap_t *snd_pcm_direct_get_chmap(snd_pcm_t *pcm); int snd_pcm_direct_set_chmap(snd_pcm_t *pcm, const snd_pcm_chmap_t *map); - +int snd_pcm_direct_slave_recover(snd_pcm_direct_t *direct); +int snd_pcm_direct_client_chk_xrun(snd_pcm_direct_t *direct, snd_pcm_t *pcm); int snd_timer_async(snd_timer_t *timer, int sig, pid_t pid); struct timespec snd_pcm_hw_fast_tstamp(snd_pcm_t *pcm); diff --git a/src/pcm/pcm_dmix.c b/src/pcm/pcm_dmix.c index 0ab7323..5c3eb2d 100644 --- a/src/pcm/pcm_dmix.c +++ b/src/pcm/pcm_dmix.c @@ -396,6 +396,7 @@ static int snd_pcm_dmix_sync_ptr0(snd_pcm_t *pcm, snd_pcm_uframes_t slave_hw_ptr snd_pcm_direct_t *dmix = pcm->private_data; snd_pcm_uframes_t old_slave_hw_ptr, avail; snd_pcm_sframes_t diff; + int err; old_slave_hw_ptr = dmix->slave_hw_ptr; dmix->slave_hw_ptr = *dmix->spcm->hw.ptr; @@ -434,15 +435,21 @@ static int snd_pcm_dmix_sync_ptr0(snd_pcm_t *pcm, snd_pcm_uframes_t slave_hw_ptr static int snd_pcm_dmix_sync_ptr(snd_pcm_t *pcm) { snd_pcm_direct_t *dmix = pcm->private_data; + int err; switch (snd_pcm_state(dmix->spcm)) { case SND_PCM_STATE_DISCONNECTED: dmix->state = SND_PCM_STATE_DISCONNECTED; return -ENODEV; + case SND_PCM_STATE_XRUN: + if ((err = snd_pcm_direct_slave_recover(dmix)) <0) + return err; + break; default: break; } - + if (snd_pcm_direct_client_chk_xrun(dmix, pcm)) + return -EPIPE; if (dmix->slowptr) snd_pcm_hwsync(dmix->spcm); @@ -828,12 +835,16 @@ static snd_pcm_sframes_t snd_pcm_dmix_mmap_commit(snd_pcm_t *pcm, switch (snd_pcm_state(dmix->spcm)) { case SND_PCM_STATE_XRUN: - return -EPIPE; + if ((err = snd_pcm_direct_slave_recover(dmix)) < 0) + return err; + break; case SND_PCM_STATE_SUSPENDED: return -ESTRPIPE; default: break; } + if (snd_pcm_direct_client_chk_xrun(dmix, pcm)) + return -EPIPE; if (! size) return 0; snd_pcm_mmap_appl_forward(pcm, size); @@ -841,8 +852,10 @@ static snd_pcm_sframes_t snd_pcm_dmix_mmap_commit(snd_pcm_t *pcm, if ((err = snd_pcm_dmix_start_timer(pcm, dmix)) < 0) return err; } else if (dmix->state == SND_PCM_STATE_RUNNING || - dmix->state == SND_PCM_STATE_DRAINING) - snd_pcm_dmix_sync_ptr(pcm); + dmix->state == SND_PCM_STATE_DRAINING) { + if (( err = snd_pcm_dmix_sync_ptr(pcm)) < 0) + return err; + } if (dmix->state == SND_PCM_STATE_RUNNING || dmix->state == SND_PCM_STATE_DRAINING) { /* ok, we commit the changes after the validation of area */ @@ -858,10 +871,13 @@ static snd_pcm_sframes_t snd_pcm_dmix_mmap_commit(snd_pcm_t *pcm, static snd_pcm_sframes_t snd_pcm_dmix_avail_update(snd_pcm_t *pcm) { snd_pcm_direct_t *dmix = pcm->private_data; + int err; if (dmix->state == SND_PCM_STATE_RUNNING || - dmix->state == SND_PCM_STATE_DRAINING) - snd_pcm_dmix_sync_ptr(pcm); + dmix->state == SND_PCM_STATE_DRAINING) { + if (( err = snd_pcm_dmix_sync_ptr(pcm)) < 0) + return err; + } return snd_pcm_mmap_playback_avail(pcm); } diff --git a/src/pcm/pcm_dshare.c b/src/pcm/pcm_dshare.c index a1fed5d..707e2c2 100644 --- a/src/pcm/pcm_dshare.c +++ b/src/pcm/pcm_dshare.c @@ -202,15 +202,21 @@ static int snd_pcm_dshare_sync_ptr0(snd_pcm_t *pcm, snd_pcm_uframes_t slave_hw_p static int snd_pcm_dshare_sync_ptr(snd_pcm_t *pcm) { snd_pcm_direct_t *dshare = pcm->private_data; + int err; switch (snd_pcm_state(dshare->spcm)) { case SND_PCM_STATE_DISCONNECTED: dshare->state = SNDRV_PCM_STATE_DISCONNECTED; return -ENODEV; + case SND_PCM_STATE_XRUN: + if ((err = snd_pcm_direct_slave_recover(dshare)) <0) + return err; + break; default: break; } - + if (snd_pcm_direct_client_chk_xrun(dshare, pcm)) + return -EPIPE; if (dshare->slowptr) snd_pcm_hwsync(dshare->spcm); @@ -516,12 +522,16 @@ static snd_pcm_sframes_t snd_pcm_dshare_mmap_commit(snd_pcm_t *pcm, switch (snd_pcm_state(dshare->spcm)) { case SND_PCM_STATE_XRUN: - return -EPIPE; + if ((err = snd_pcm_direct_slave_recover(dshare)) < 0) + return err; + break; case SND_PCM_STATE_SUSPENDED: return -ESTRPIPE; default: break; } + if (snd_pcm_direct_client_chk_xrun(dshare, pcm)) + return -EPIPE; if (! size) return 0; snd_pcm_mmap_appl_forward(pcm, size); @@ -529,8 +539,10 @@ static snd_pcm_sframes_t snd_pcm_dshare_mmap_commit(snd_pcm_t *pcm, if ((err = snd_pcm_dshare_start_timer(dshare)) < 0) return err; } else if (dshare->state == SND_PCM_STATE_RUNNING || - dshare->state == SND_PCM_STATE_DRAINING) - snd_pcm_dshare_sync_ptr(pcm); + dshare->state == SND_PCM_STATE_DRAINING) { + if ((err = snd_pcm_dshare_sync_ptr(pcm)) < 0) + return err; + } if (dshare->state == SND_PCM_STATE_RUNNING || dshare->state == SND_PCM_STATE_DRAINING) { /* ok, we commit the changes after the validation of area */ @@ -546,10 +558,13 @@ static snd_pcm_sframes_t snd_pcm_dshare_mmap_commit(snd_pcm_t *pcm, static snd_pcm_sframes_t snd_pcm_dshare_avail_update(snd_pcm_t *pcm) { snd_pcm_direct_t *dshare = pcm->private_data; + int err; if (dshare->state == SND_PCM_STATE_RUNNING || - dshare->state == SND_PCM_STATE_DRAINING) - snd_pcm_dshare_sync_ptr(pcm); + dshare->state == SND_PCM_STATE_DRAINING) { + if ((err = snd_pcm_dshare_sync_ptr(pcm)) < 0) + return err; + } return snd_pcm_mmap_playback_avail(pcm); } diff --git a/src/pcm/pcm_dsnoop.c b/src/pcm/pcm_dsnoop.c index 85f0ff4..512e957 100644 --- a/src/pcm/pcm_dsnoop.c +++ b/src/pcm/pcm_dsnoop.c @@ -132,14 +132,21 @@ static int snd_pcm_dsnoop_sync_ptr(snd_pcm_t *pcm) snd_pcm_direct_t *dsnoop = pcm->private_data; snd_pcm_uframes_t slave_hw_ptr, old_slave_hw_ptr, avail; snd_pcm_sframes_t diff; + int err; switch (snd_pcm_state(dsnoop->spcm)) { case SND_PCM_STATE_DISCONNECTED: dsnoop->state = SNDRV_PCM_STATE_DISCONNECTED; return -ENODEV; + case SND_PCM_STATE_XRUN: + if ((err = snd_pcm_direct_slave_recover(dsnoop)) <0) + return err; + break; default: break; } + if (snd_pcm_direct_client_chk_xrun(dsnoop, pcm)) + return -EPIPE; if (dsnoop->slowptr) snd_pcm_hwsync(dsnoop->spcm); old_slave_hw_ptr = dsnoop->slave_hw_ptr; @@ -410,12 +417,16 @@ static snd_pcm_sframes_t snd_pcm_dsnoop_mmap_commit(snd_pcm_t *pcm, switch (snd_pcm_state(dsnoop->spcm)) { case SND_PCM_STATE_XRUN: - return -EPIPE; + if ((err = snd_pcm_direct_slave_recover(dsnoop)) <0) + return err; + break; case SND_PCM_STATE_SUSPENDED: return -ESTRPIPE; default: break; } + if (snd_pcm_direct_client_chk_xrun(dsnoop, pcm)) + return -EPIPE; if (dsnoop->state == SND_PCM_STATE_RUNNING) { err = snd_pcm_dsnoop_sync_ptr(pcm); if (err < 0) -- 1.7.9.5

2 1

[alsa-devel] [PATCH 1/6] alsa-lib:pcm: allow users to configure different period sizes.
by sutar.mounesh＠gmail.com 02 Jan '17

02 Jan '17

From: Joshua Frkuska joshua_frkuska(a)mentor.com This patch allows the effective period size to be a multiple of the slave-pcm period size. Allowing only exact multiple of original period size is achieved by borrowing code from the kernel hwrules implemention. This patch is intended to save cpu workload when for example, the slave operates with very small periods but a user does not need that small periods. This feature is enabled by default and can be disabled by adding config option 'var_periodsize 0'. Signed-off-by: Alexander Jahn <ajahn(a)de.adit-jv.com> Signed-off-by: Andreas Pape <apape(a)de.adit-jv.com> diff --git a/src/pcm/pcm_direct.c b/src/pcm/pcm_direct.c index 6434983..8f42b19 100644 --- a/src/pcm/pcm_direct.c +++ b/src/pcm/pcm_direct.c @@ -660,6 +660,28 @@ static int hw_param_interval_refine_minmax(snd_pcm_hw_params_t *params, return hw_param_interval_refine_one(params, var, &t); } +/* this code is used 'as-is' from the alsa kernel code */ +static int snd_interval_step(struct snd_interval *i, unsigned int min, unsigned int step) +{ + unsigned int n; + int changed = 0; + n = (i->min - min) % step; + if (n != 0 || i->openmin) { + i->min += step - n; + changed = 1; + } + n = (i->max - min) % step; + if (n != 0 || i->openmax) { + i->max -= n; + changed = 1; + } + if (snd_interval_checkempty(i)) { + i->empty = 1; + return -EINVAL; + } + return changed; +} + #undef REFINE_DEBUG int snd_pcm_direct_hw_refine(snd_pcm_t *pcm, snd_pcm_hw_params_t *params) @@ -710,15 +732,16 @@ int snd_pcm_direct_hw_refine(snd_pcm_t *pcm, snd_pcm_hw_params_t *params) &dshare->shmptr->hw.rate); if (err < 0) return err; - err = hw_param_interval_refine_one(params, SND_PCM_HW_PARAM_PERIOD_SIZE, - &dshare->shmptr->hw.period_size); - if (err < 0) - return err; - err = hw_param_interval_refine_one(params, SND_PCM_HW_PARAM_PERIOD_TIME, - &dshare->shmptr->hw.period_time); - if (err < 0) - return err; + if (dshare->max_periods < 0) { + err = hw_param_interval_refine_one(params, SND_PCM_HW_PARAM_PERIOD_SIZE, + &dshare->shmptr->hw.period_size); + if (err < 0) + return err; + err = hw_param_interval_refine_one(params, SND_PCM_HW_PARAM_PERIOD_TIME, + &dshare->shmptr->hw.period_time); + if (err < 0) + return err; err = hw_param_interval_refine_one(params, SND_PCM_HW_PARAM_BUFFER_SIZE, &dshare->shmptr->hw.buffer_size); if (err < 0) @@ -730,11 +753,38 @@ int snd_pcm_direct_hw_refine(snd_pcm_t *pcm, snd_pcm_hw_params_t *params) } else if (params->rmask & ((1<<SND_PCM_HW_PARAM_PERIODS)| (1<<SND_PCM_HW_PARAM_BUFFER_BYTES)| (1<<SND_PCM_HW_PARAM_BUFFER_SIZE)| - (1<<SND_PCM_HW_PARAM_BUFFER_TIME))) { + (1<<SND_PCM_HW_PARAM_BUFFER_TIME)| + (1<<SND_PCM_HW_PARAM_PERIOD_TIME)| + (1<<SND_PCM_HW_PARAM_PERIOD_SIZE)| + (1<<SND_PCM_HW_PARAM_PERIOD_BYTES))) { + snd_interval_t period_size = dshare->shmptr->hw.period_size; + snd_interval_t period_time = dshare->shmptr->hw.period_time; int changed; unsigned int max_periods = dshare->max_periods; if (max_periods < 2) max_periods = dshare->slave_buffer_size / dshare->slave_period_size; + + /*make sure buffer size does not exceed slave buffer size*/ + err = hw_param_interval_refine_minmax(params, SND_PCM_HW_PARAM_BUFFER_SIZE, + 2 * dshare->slave_period_size, dshare->slave_buffer_size); + if (err < 0) + return err; + if (dshare->var_periodsize) { + /*more tolerant settings...*/ + if((dshare->shmptr->hw.buffer_size.max / 2) > period_size.max) + period_size.max = dshare->shmptr->hw.buffer_size.max / 2; + if((dshare->shmptr->hw.buffer_time.max / 2) > period_time.max) + period_time.max = dshare->shmptr->hw.buffer_time.max / 2; + } + + err = hw_param_interval_refine_one(params, SND_PCM_HW_PARAM_PERIOD_SIZE, + &period_size); + if (err < 0) + return err; + err = hw_param_interval_refine_one(params, SND_PCM_HW_PARAM_PERIOD_TIME, + &period_time); + if (err < 0) + return err; do { changed = 0; err = hw_param_interval_refine_minmax(params, SND_PCM_HW_PARAM_PERIODS, @@ -746,8 +796,16 @@ int snd_pcm_direct_hw_refine(snd_pcm_t *pcm, snd_pcm_hw_params_t *params) if (err < 0) return err; changed |= err; + err = snd_interval_step(hw_param_interval(params, SND_PCM_HW_PARAM_PERIOD_SIZE), + 0, dshare->slave_period_size); + if (err < 0) + return err; + changed |= err; + if (err) + params->rmask |= (1<<SND_PCM_HW_PARAM_PERIOD_SIZE); } while (changed); } + dshare->timer_ticks = hw_param_interval(params, SND_PCM_HW_PARAM_PERIOD_SIZE)->max/dshare->slave_period_size; params->info = dshare->shmptr->s.info; #ifdef REFINE_DEBUG snd_output_puts(log, "DMIX REFINE (end):\n"); @@ -1183,6 +1241,7 @@ int snd_pcm_direct_initialize_poll_fd(snd_pcm_direct_t *dmix) dmix->tread = 1; dmix->timer_need_poll = 0; + dmix->timer_ticks = 1; ret = snd_pcm_info(dmix->spcm, &info); if (ret < 0) { SNDERR("unable to info for slave pcm"); @@ -1366,7 +1425,7 @@ int snd_pcm_direct_set_timer_params(snd_pcm_direct_t *dmix) snd_timer_params_set_auto_start(&params, 1); if (dmix->type != SND_PCM_TYPE_DSNOOP) snd_timer_params_set_early_event(&params, 1); - snd_timer_params_set_ticks(&params, 1); + snd_timer_params_set_ticks(&params, dmix->timer_ticks); if (dmix->tread) { filter = (1<<SND_TIMER_EVENT_TICK) | dmix->timer_events; @@ -1656,6 +1715,7 @@ int snd_pcm_direct_parse_open_conf(snd_config_t *root, snd_config_t *conf, rec->ipc_gid = -1; rec->slowptr = 1; rec->max_periods = 0; + rec->var_periodsize = 1; /* read defaults */ if (snd_config_search(root, "defaults.pcm.dmix_max_periods", &n) >= 0) { @@ -1762,6 +1822,13 @@ int snd_pcm_direct_parse_open_conf(snd_config_t *root, snd_config_t *conf, rec->max_periods = val; continue; } + if (strcmp(id, "var_periodsize") == 0) { + err = snd_config_get_bool(n); + if (err < 0) + return err; + rec->var_periodsize = err; + continue; + } SNDERR("Unknown field %s", id); return -EINVAL; } diff --git a/src/pcm/pcm_direct.h b/src/pcm/pcm_direct.h index 611ad29..91e816c 100644 --- a/src/pcm/pcm_direct.h +++ b/src/pcm/pcm_direct.h @@ -147,12 +147,14 @@ struct snd_pcm_direct { int tread: 1; int timer_need_poll: 1; unsigned int timer_events; + unsigned int timer_ticks; int server_fd; pid_t server_pid; snd_timer_t *timer; /* timer used as poll_fd */ int interleaved; /* we have interleaved buffer */ int slowptr; /* use slow but more precise ptr updates */ int max_periods; /* max periods (-1 = fixed periods, 0 = max buffer size) */ + int var_periodsize; /* allow variable period size if max_periods is != -1*/ unsigned int channels; /* client's channels */ unsigned int *bindings; union { @@ -326,6 +328,7 @@ struct snd_pcm_direct_open_conf { int ipc_gid; int slowptr; int max_periods; + int var_periodsize; snd_config_t *slave; snd_config_t *bindings; }; diff --git a/src/pcm/pcm_dmix.c b/src/pcm/pcm_dmix.c index 825677f..0ab7323 100644 --- a/src/pcm/pcm_dmix.c +++ b/src/pcm/pcm_dmix.c @@ -1040,6 +1040,7 @@ int snd_pcm_dmix_open(snd_pcm_t **pcmp, const char *name, dmix->state = SND_PCM_STATE_OPEN; dmix->slowptr = opts->slowptr; dmix->max_periods = opts->max_periods; + dmix->var_periodsize = opts->var_periodsize; dmix->sync_ptr = snd_pcm_dmix_sync_ptr; retry: diff --git a/src/pcm/pcm_dshare.c b/src/pcm/pcm_dshare.c index 29cd6c6..a1fed5d 100644 --- a/src/pcm/pcm_dshare.c +++ b/src/pcm/pcm_dshare.c @@ -725,6 +725,7 @@ int snd_pcm_dshare_open(snd_pcm_t **pcmp, const char *name, dshare->state = SND_PCM_STATE_OPEN; dshare->slowptr = opts->slowptr; dshare->max_periods = opts->max_periods; + dshare->var_periodsize = opts->var_periodsize; dshare->sync_ptr = snd_pcm_dshare_sync_ptr; retry: diff --git a/src/pcm/pcm_dsnoop.c b/src/pcm/pcm_dsnoop.c index 1aedf3c..85f0ff4 100644 --- a/src/pcm/pcm_dsnoop.c +++ b/src/pcm/pcm_dsnoop.c @@ -606,6 +606,7 @@ int snd_pcm_dsnoop_open(snd_pcm_t **pcmp, const char *name, dsnoop->state = SND_PCM_STATE_OPEN; dsnoop->slowptr = opts->slowptr; dsnoop->max_periods = opts->max_periods; + dsnoop->var_periodsize = opts->var_periodsize; dsnoop->sync_ptr = snd_pcm_dsnoop_sync_ptr; retry: -- 1.7.9.5

2 1

[alsa-devel] [PATCH 0/3] ASoC: tlv320aic3x: Pop noise reductions
by Peter Ujfalusi 02 Jan '17

02 Jan '17

Hi, When we don't have any active component between the output of the codec and the speaker/jack the codec emmits pop noise due to several factors: - RESET register is toggled by regcache_sync (patch 1) - After regcache_sync we need to wait a bit to allow the codec to settle down: the codec will be soft reseted when it is powered down, so when we power it on and run regcache_sync, we need to wait for the flushed settings to be valid. (patch 2) - The HP ouptut power-down mode can be changed to be driven weakly to VCM, this can reduce the pop noise as well, with slight increase in powere consumption. These factors were not taken care as in n810 and n900 we have external active amplifiers and they will remove any artifacts coming from the codec, but on boards w/o active external components they can reach the output. Regards, Peter --- Peter Ujfalusi (3): ASoC: tlv320aic3x: Mark the RESET register as volatile ASoC: tlv320aic3x: Add delay after power on and register sync ASoC: tlv320aic3x: Add controls for selecting HP power down modes sound/soc/codecs/tlv320aic3x.c | 29 +++++++++++++++++++++++++++++ 1 file changed, 29 insertions(+) -- 2.11.0

3 9

[alsa-devel] snd_hda_intel/snd_hda_codec_hdmi module load/unload race
by Imre Deak 02 Jan '17

02 Jan '17

Hi, I got the trace below while trying to unload (unbind) snd_hda_intel, while its still loading the HDMI codec driver. IIUC what happens is: Task1 Task2 Task3 modprobe snd_hda_intel schedule(azx_probe_work) unbind snd_hda_intel via sysfs device_release_driver() device_lock(snd_hda_intel) azx_remove() cancel_work_sync(azx_probe_work) azx_probe_work() request_module(snd-hda-codec-hdmi) hdmi_driver_init() __driver_attach() device_lock(snd_hda_intel) Deadlock, since azx_probe_work() will never finish and the snd_hda_intel device lock will never get released. --Imre [ 246.805268] INFO: task drv_module_relo:1684 blocked for more than 120 seconds. [ 246.805834] Tainted: G W 4.9.0-CI-CI_DRM_1946+ #158 [ 246.806284] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. [ 246.806779] drv_module_relo D 0 1684 1597 0x00000000 [ 246.807171] ffff8802703cbac0 ffff8802643926c0 000000000000479d ffff8802713b0040 [ 246.807706] ffff88027fc98118 ffffc90003cffa58 ffffffff8181d429 0000000000000046 [ 246.808217] ffffffff827d2470 0000000000000001 ffff88027fc98118 ffff88027fc98100 [ 246.808739] Call Trace: [ 246.808945] [<ffffffff8181d429>] ? __schedule+0x2c9/0xa30 [ 246.809281] [<ffffffff8181dbcb>] schedule+0x3b/0x90 [ 246.809618] [<ffffffff8182227b>] schedule_timeout+0x25b/0x4c0 [ 246.810021] [<ffffffff810d6c9f>] ? mark_held_locks+0x6f/0xa0 [ 246.810370] [<ffffffff81823987>] ? _raw_spin_unlock_irq+0x27/0x50 [ 246.810759] [<ffffffff810d6df2>] ? trace_hardirqs_on_caller+0x122/0x1b0 [ 246.811963] [<ffffffff8181e730>] wait_for_common+0xf0/0x160 [ 246.812463] [<ffffffff810af5d0>] ? wake_up_q+0x70/0x70 [ 246.812795] [<ffffffff8181e7b8>] wait_for_completion+0x18/0x20 [ 246.813389] [<ffffffff8109b9e3>] flush_work+0x203/0x2e0 [ 246.813714] [<ffffffff8109ba86>] ? flush_work+0x2a6/0x2e0 [ 246.814088] [<ffffffff81099410>] ? flush_workqueue_prep_pwqs+0x220/0x220 [ 246.814790] [<ffffffff8109d290>] __cancel_work_timer+0x110/0x1d0 [ 246.815346] [<ffffffff810d6df2>] ? trace_hardirqs_on_caller+0x122/0x1b0 [ 246.815742] [<ffffffff810d6e8d>] ? trace_hardirqs_on+0xd/0x10 [ 246.816096] [<ffffffff8109d35b>] cancel_work_sync+0xb/0x10 [ 246.816566] [<ffffffffa00d4fa4>] azx_remove+0x24/0x30 [snd_hda_intel] [ 246.816956] [<ffffffff81485864>] pci_device_remove+0x34/0xb0 [ 246.817373] [<ffffffff81584cfc>] __device_release_driver+0x9c/0x150 [ 246.817717] [<ffffffff81584dd0>] device_release_driver+0x20/0x30 [ 246.818077] [<ffffffff815836aa>] unbind_store+0x10a/0x150 [ 246.818549] [<ffffffff81582d70>] drv_attr_store+0x20/0x30 [ 246.818885] [<ffffffff812775a0>] sysfs_kf_write+0x40/0x50 [ 246.819290] [<ffffffff812768e0>] kernfs_fop_write+0x130/0x1b0 [ 246.819639] [<ffffffff811f6383>] __vfs_write+0x23/0x120 [ 246.819963] [<ffffffff810f48c0>] ? rcu_read_lock_sched_held+0x70/0x80 [ 246.820476] [<ffffffff810f4caa>] ? rcu_sync_lockdep_assert+0x2a/0x50 [ 246.820859] [<ffffffff811f9a68>] ? __sb_start_write+0xf8/0x200 [ 246.821292] [<ffffffff811f7129>] ? vfs_write+0x169/0x1b0 [ 246.821597] [<ffffffff811f7073>] vfs_write+0xb3/0x1b0 [ 246.821907] [<ffffffff810d6df2>] ? trace_hardirqs_on_caller+0x122/0x1b0 [ 246.822405] [<ffffffff811f83b4>] SyS_write+0x44/0xb0 [ 246.822684] [<ffffffff8182412e>] entry_SYSCALL_64_fastpath+0x1c/0xb1 [ 246.823064] Showing all locks held in the system: [ 246.823568] 2 locks held by kworker/u8:0/6: [ 246.823817] #0: ("events_unbound"){.+.+.+}, at: [<ffffffff8109c1f6>] process_one_work+0x166/0x6b0 [ 246.824400] #1: ((&sub_info->work)){+.+.+.}, at: [<ffffffff8109c1f6>] process_one_work+0x166/0x6b0 [ 246.824940] 2 locks held by khungtaskd/34: [ 246.825210] #0: (rcu_read_lock){......}, at: [<ffffffff811329cc>] watchdog+0x9c/0x600 [ 246.825726] #1: (tasklist_lock){.+.+..}, at: [<ffffffff810d536d>] debug_show_all_locks+0x3d/0x1a0 [ 246.826303] 2 locks held by kworker/3:1/50: [ 246.826535] #0: ("events"){++++.+}, at: [<ffffffff8109c1f6>] process_one_work+0x166/0x6b0 [ 246.827104] #1: ((&hda->probe_work)){+.+.+.}, at: [<ffffffff8109c1f6>] process_one_work+0x166/0x6b0 [ 246.827706] 5 locks held by drv_module_relo/1684: [ 246.827963] #0: (sb_writers#6){.+.+.+}, at: [<ffffffff811f7129>] vfs_write+0x169/0x1b0 [ 246.828507] #1: (&of->mutex){+.+.+.}, at: [<ffffffff812768ac>] kernfs_fop_write+0xfc/0x1b0 [ 246.829083] #2: (s_active#186){++++.+}, at: [<ffffffff812768b4>] kernfs_fop_write+0x104/0x1b0 [ 246.829663] #3: (&dev->mutex){......}, at: [<ffffffff8158369a>] unbind_store+0xfa/0x150 [ 246.830204] #4: (&dev->mutex){......}, at: [<ffffffff81584dc8>] device_release_driver+0x18/0x30 [ 246.830772] 1 lock held by modprobe/1685: [ 246.830992] #0: (&dev->mutex){......}, at: [<ffffffff8158561a>] __driver_attach+0x5a/0xe0 [ 246.831575] 2 locks held by bash/1727: [ 246.831782] #0: (&tty->ldisc_sem){++++.+}, at: [<ffffffff81822d3d>] ldsem_down_read+0x2d/0x40 [ 246.832344] #1: (&ldata->atomic_read_lock){+.+...}, at: [<ffffffff81506529>] n_tty_read+0xa9/0x880 [ 246.833050] ============================================= [ 246.833436] INFO: task modprobe:1685 blocked for more than 120 seconds. [ 246.833829] Tainted: G W 4.9.0-CI-CI_DRM_1946+ #158 [ 246.834221] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. [ 246.834680] modprobe D 0 1685 6 0x00000000 [ 246.835029] ffff8802703c9a40 ffff8802760e4d40 00000000000091ba ffff8802559e26c0 [ 246.835522] ffff88027fd18118 ffffc900032bbb70 ffffffff8181d429 0000000000000000 [ 246.835967] ffffc900032bbb98 0000000000000292 ffff88027fd18118 ffff88027fd18100 [ 246.836462] Call Trace: [ 246.836628] [<ffffffff8181d429>] ? __schedule+0x2c9/0xa30 [ 246.836924] [<ffffffff8181dbcb>] schedule+0x3b/0x90 [ 246.837243] [<ffffffff8181e043>] schedule_preempt_disabled+0x13/0x20 [ 246.837629] [<ffffffff8181ecbc>] mutex_lock_nested+0x16c/0x400 [ 246.837946] [<ffffffff8158561a>] ? __driver_attach+0x5a/0xe0 [ 246.838298] [<ffffffff815855c0>] ? driver_probe_device+0x430/0x430 [ 246.838676] [<ffffffff8158561a>] __driver_attach+0x5a/0xe0 [ 246.839008] [<ffffffff815855c0>] ? driver_probe_device+0x430/0x430 [ 246.839354] [<ffffffff81582fc1>] bus_for_each_dev+0x61/0xa0 [ 246.839704] [<ffffffff81584b89>] driver_attach+0x19/0x20 [ 246.840029] [<ffffffff815846cb>] bus_add_driver+0x1fb/0x270 [ 246.840346] [<ffffffffa014a000>] ? 0xffffffffa014a000 [ 246.840645] [<ffffffff8158629b>] driver_register+0x5b/0xd0 [ 246.840977] [<ffffffffa048f0d5>] __hda_codec_driver_register+0x55/0x60 [snd_hda_codec] [ 246.841417] [<ffffffffa014a01e>] hdmi_driver_init+0x1e/0x20 [snd_hda_codec_hdmi] [ 246.841861] [<ffffffff81000428>] do_one_initcall+0x38/0x150 [ 246.842177] [<ffffffff810f48c0>] ? rcu_read_lock_sched_held+0x70/0x80 [ 246.842566] [<ffffffff811d7804>] ? kmem_cache_alloc_trace+0x274/0x2e0 [ 246.842951] [<ffffffff81176161>] ? do_init_module+0x22/0x1d1 [ 246.843307] [<ffffffff81176199>] do_init_module+0x5a/0x1d1 [ 246.843617] [<ffffffff8111cb15>] load_module+0x2275/0x27d0 [ 246.843951] [<ffffffff81119e60>] ? symbol_put_addr+0x60/0x60 [ 246.844306] [<ffffffff811fd7d5>] ? kernel_read_file+0x105/0x190 [ 246.844637] [<ffffffff8111d29c>] SyS_finit_module+0xbc/0xf0 [ 246.844979] [<ffffffff8182412e>] entry_SYSCALL_64_fastpath+0x1c/0xb1 [ 246.845368] Showing all locks held in the system: [ 246.845745] 2 locks held by kworker/u8:0/6: [ 246.845974] #0: ("events_unbound"){.+.+.+}, at: [<ffffffff8109c1f6>] process_one_work+0x166/0x6b0 [ 246.846553] #1: ((&sub_info->work)){+.+.+.}, at: [<ffffffff8109c1f6>] process_one_work+0x166/0x6b0 [ 246.847133] 2 locks held by khungtaskd/34: [ 246.847399] #0: (rcu_read_lock){......}, at: [<ffffffff811329cc>] watchdog+0x9c/0x600 [ 246.847913] #1: (tasklist_lock){.+.+..}, at: [<ffffffff810d536d>] debug_show_all_locks+0x3d/0x1a0 [ 246.848488] 2 locks held by kworker/3:1/50: [ 246.848747] #0: ("events"){++++.+}, at: [<ffffffff8109c1f6>] process_one_work+0x166/0x6b0 [ 246.849286] #1: ((&hda->probe_work)){+.+.+.}, at: [<ffffffff8109c1f6>] process_one_work+0x166/0x6b0 [ 246.849875] 1 lock held by in:imklog/491: [ 246.850094] #0: (&f->f_pos_lock){+.+.+.}, at: [<ffffffff8121854e>] __fdget_pos+0x3e/0x50 [ 246.850657] 5 locks held by drv_module_relo/1684: [ 246.850931] #0: (sb_writers#6){.+.+.+}, at: [<ffffffff811f7129>] vfs_write+0x169/0x1b0 [ 246.851474] #1: (&of->mutex){+.+.+.}, at: [<ffffffff812768ac>] kernfs_fop_write+0xfc/0x1b0 [ 246.852013] #2: (s_active#186){++++.+}, at: [<ffffffff812768b4>] kernfs_fop_write+0x104/0x1b0 [ 246.852577] #3: (&dev->mutex){......}, at: [<ffffffff8158369a>] unbind_store+0xfa/0x150 [ 246.853103] #4: (&dev->mutex){......}, at: [<ffffffff81584dc8>] device_release_driver+0x18/0x30 [ 246.853714] 1 lock held by modprobe/1685: [ 246.853934] #0: (&dev->mutex){......}, at: [<ffffffff8158561a>] __driver_attach+0x5a/0xe0 [ 246.854473] 2 locks held by bash/1727: [ 246.854713] #0: (&tty->ldisc_sem){++++.+}, at: [<ffffffff81822d3d>] ldsem_down_read+0x2d/0x40 [ 246.855272] #1: (&ldata->atomic_read_lock){+.+...}, at: [<ffffffff81506529>] n_tty_read+0xa9/0x880 Stack trace of kworker/50: [<ffffffff810979e3>] call_usermodehelper_exec+0x133/0x1d0 [<ffffffff81097c0f>] __request_module+0x18f/0x360 [<ffffffffa048f182>] request_codec_module+0x52/0x60 [snd_hda_codec] [<ffffffffa048f21d>] snd_hda_codec_configure+0x5d/0x170 [snd_hda_codec] [<ffffffffa0499693>] azx_codec_configure+0x23/0x40 [snd_hda_codec] [<ffffffffa00d6700>] azx_probe_work+0x5d0/0x8d0 [snd_hda_intel] [<ffffffff8109c27c>] process_one_work+0x1ec/0x6b0 [<ffffffff8109c789>] worker_thread+0x49/0x490 [<ffffffff810a2aab>] kthread+0xeb/0x110 [<ffffffff818243a7>] ret_from_fork+0x27/0x40 [<ffffffffffffffff>] 0xffffffffffffffff

3 3

[alsa-devel] [PATCH 0/5] Runtime PM support for Exynos I2S driver
by Marek Szyprowski 31 Dec '16

31 Dec '16

Exynos I2S audio controller was missing proper runtime PM implementation and this patchset addresses this issue. This patchset is a another step to add support for audio power domain on Exynos5 SoCs. Audio power domain on Exynos5 SoCs contains following hardware modules: 1. clock controller 2. pin controller 3. PL330 DMA controller 4. I2S audio controller The other patches related to enabling full support for audio power domain can be found here: 1. PL330 ADMA controller non-irqsafe runtime PM: https://www.spinics.net/lists/arm-kernel/msg550008.html 2. Runtime PM for clock controllers (Exynos Audio subsystem will be added in v4 soon): https://www.spinics.net/lists/arm-kernel/msg538122.html 3. Runtime PM support for pin controller: https://www.spinics.net/lists/linux-samsung-soc/msg56674.html Patches are based on linux-next from 2016.12.24. Tested on Odroid U3 (Exynos4412), Odroid XU3 (Exynos5422) and TM2 (Exynos5433) boards. Best regards Marek Szyprowski Samsung R&D Institute Poland Patch summary: Marek Szyprowski (5): ASoC: samsung: i2s: Remove virtual device for secondary DAI ASoC: samsung: i2s: Ensure proper runtime PM state of I2S device ASoC: samsung: i2s: Move saving and restoring regs to runtime pm operations ASoC: samsung: i2s: Let runtime PM operations to control op_clk too ASoC: samsung: i2s: Provide I2S device for registered clocks sound/soc/samsung/i2s.c | 193 +++++++++++++++++++++++------------------------- 1 file changed, 92 insertions(+), 101 deletions(-) -- 1.9.1

2 11