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

[ANNOUNCE] SOF v1.7 released
by Girdwood, Liam R 06 Apr '21

06 Apr '21

The SOF developers are pleased to announce that version 1.7 has been released and binaries available for Intel and NXP platforms[1]. The main focus of v1.7 was around adding support for new audio processing algorithms, continued Zephyr integration and general stabilisation. Release notes are here. https://github.com/thesofproject/sof/releases/tag/v1.7 Binaries are here. https://github.com/thesofproject/sof-bin Firmware issues can be reported here (and feature requests too). https://github.com/thesofproject/sof/issues and kernel issues here https://github.com/thesofproject/linux/issues The next version will be v1.8 due by the end of June. Thanks Liam [1] Preliminary support for other platforms can be enabled by using SOF as a Zephyr module (for Zephyr supported architectures and platforms). --------------------------------------------------------------------- Intel Corporation (UK) Limited Registered No. 1134945 (England) Registered Office: Pipers Way, Swindon SN3 1RJ VAT No: 860 2173 47 This e-mail and any attachments may contain confidential material for the sole use of the intended recipient(s). Any review or distribution by others is strictly prohibited. If you are not the intended recipient, please contact the sender and delete all copies.

2 1

[PATCH v2] soundwire: intel_init: test link->cdns
by Bard Liao 06 Apr '21

06 Apr '21

intel_link_probe() could return error and dev_get_drvdata() will return null in such case. So we have to test link->cdns after link->cdns = dev_get_drvdata(&ldev->auxdev.dev); Otherwise, we will meet the "kernel NULL pointer dereference" error. Signed-off-by: Bard Liao <yung-chuan.liao(a)linux.intel.com> Reviewed-by: Ranjani Sridharan <ranjani.sridharan(a)linux.intel.com> Reviewed-by: Rander Wang <rander.wang(a)intel.com> Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart(a)linux.intel.com> --- v2: - Rebase to latest code base. --- drivers/soundwire/intel_init.c | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/drivers/soundwire/intel_init.c b/drivers/soundwire/intel_init.c index 05b726cdfebc..30ce95ec2d70 100644 --- a/drivers/soundwire/intel_init.c +++ b/drivers/soundwire/intel_init.c @@ -178,6 +178,15 @@ static struct sdw_intel_ctx link->pdev = pdev; link->cdns = platform_get_drvdata(pdev); + if (!link->cdns) { + dev_err(&adev->dev, "failed to get link->cdns\n"); + /* + * 1 will be subtracted from i in the err label, but we need to call + * intel_link_dev_unregister for this ldev, so plus 1 now + */ + i++; + goto err; + } list_add_tail(&link->list, &ctx->link_list); bus = &link->cdns->bus; /* Calculate number of slaves */ -- 2.17.1

2 1

[PATCH] soundwire: qcom: handle return correctly in qcom_swrm_transport_params
by Srinivas Kandagatla 06 Apr '21

06 Apr '21

Looks like return from reg_write is set but not checked. Fix this by adding error return path. Reported-by: coverity-bot <keescook+coverity-bot(a)chromium.org> Addresses-Coverity-ID: 1503591 ("UNUSED_VALUE") Fixes: 128eaf937adb ("soundwire: qcom: add support to missing transport params") Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla(a)linaro.org> --- drivers/soundwire/qcom.c | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/drivers/soundwire/qcom.c b/drivers/soundwire/qcom.c index 5fd4a99cc8ac..348d9a46f850 100644 --- a/drivers/soundwire/qcom.c +++ b/drivers/soundwire/qcom.c @@ -731,17 +731,23 @@ static int qcom_swrm_transport_params(struct sdw_bus *bus, value |= pcfg->si; ret = ctrl->reg_write(ctrl, reg, value); + if (ret) + goto err; if (pcfg->lane_control != SWR_INVALID_PARAM) { reg = SWRM_DP_PORT_CTRL_2_BANK(params->port_num, bank); value = pcfg->lane_control; ret = ctrl->reg_write(ctrl, reg, value); + if (ret) + goto err; } if (pcfg->blk_group_count != SWR_INVALID_PARAM) { reg = SWRM_DP_BLOCK_CTRL2_BANK(params->port_num, bank); value = pcfg->blk_group_count; ret = ctrl->reg_write(ctrl, reg, value); + if (ret) + goto err; } if (pcfg->hstart != SWR_INVALID_PARAM @@ -755,11 +761,15 @@ static int qcom_swrm_transport_params(struct sdw_bus *bus, ret = ctrl->reg_write(ctrl, reg, value); } + if (ret) + goto err; + if (pcfg->bp_mode != SWR_INVALID_PARAM) { reg = SWRM_DP_BLOCK_CTRL3_BANK(params->port_num, bank); ret = ctrl->reg_write(ctrl, reg, pcfg->bp_mode); } +err: return ret; } -- 2.21.0

3 2

[PATCH v2] soundwire: qcom: wait for fifo space to be available before read/write
by Srinivas Kandagatla 06 Apr '21

06 Apr '21

If we write registers very fast we can endup in a situation where some of the writes will be dropped without any notice. So wait for the fifo space to be available before reading/writing the soundwire registers. Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla(a)linaro.org> --- Changes since v1: merged some of the loop code to make it simple as suggested by Pierre updated error code and comments as suggested by Vinod drivers/soundwire/qcom.c | 66 ++++++++++++++++++++++++++++++++++++++++ 1 file changed, 66 insertions(+) diff --git a/drivers/soundwire/qcom.c b/drivers/soundwire/qcom.c index 6affa3cd4039..5fd4a99cc8ac 100644 --- a/drivers/soundwire/qcom.c +++ b/drivers/soundwire/qcom.c @@ -24,6 +24,8 @@ #define SWRM_COMP_CFG_IRQ_LEVEL_OR_PULSE_MSK BIT(1) #define SWRM_COMP_CFG_ENABLE_MSK BIT(0) #define SWRM_COMP_PARAMS 0x100 +#define SWRM_COMP_PARAMS_WR_FIFO_DEPTH GENMASK(14, 10) +#define SWRM_COMP_PARAMS_RD_FIFO_DEPTH GENMASK(19, 15) #define SWRM_COMP_PARAMS_DOUT_PORTS_MASK GENMASK(4, 0) #define SWRM_COMP_PARAMS_DIN_PORTS_MASK GENMASK(9, 5) #define SWRM_INTERRUPT_STATUS 0x200 @@ -51,6 +53,8 @@ #define SWRM_CMD_FIFO_CMD 0x308 #define SWRM_CMD_FIFO_FLUSH 0x1 #define SWRM_CMD_FIFO_STATUS 0x30C +#define SWRM_RD_CMD_FIFO_CNT_MASK GENMASK(20, 16) +#define SWRM_WR_CMD_FIFO_CNT_MASK GENMASK(12, 8) #define SWRM_CMD_FIFO_CFG_ADDR 0x314 #define SWRM_CONTINUE_EXEC_ON_CMD_IGNORE BIT(31) #define SWRM_RD_WR_CMD_RETRIES 0x7 @@ -104,6 +108,7 @@ #define SWR_BROADCAST_CMD_ID 0x0F #define SWR_MAX_CMD_ID 14 #define MAX_FIFO_RD_RETRY 3 +#define SWR_OVERFLOW_RETRY_COUNT 30 struct qcom_swrm_port_config { u8 si; @@ -147,6 +152,8 @@ struct qcom_swrm_ctrl { int (*reg_read)(struct qcom_swrm_ctrl *ctrl, int reg, u32 *val); int (*reg_write)(struct qcom_swrm_ctrl *ctrl, int reg, int val); u32 slave_status; + u32 wr_fifo_depth; + u32 rd_fifo_depth; }; struct qcom_swrm_data { @@ -238,6 +245,55 @@ static u32 swrm_get_packed_reg_val(u8 *cmd_id, u8 cmd_data, return val; } +static int swrm_wait_for_rd_fifo_avail(struct qcom_swrm_ctrl *swrm) +{ + u32 fifo_outstanding_data, value; + int fifo_retry_count = SWR_OVERFLOW_RETRY_COUNT; + + do { + /* Check for fifo underflow during read */ + swrm->reg_read(swrm, SWRM_CMD_FIFO_STATUS, &value); + fifo_outstanding_data = FIELD_GET(SWRM_RD_CMD_FIFO_CNT_MASK, value); + + /* Check if read data is available in read fifo */ + if (fifo_outstanding_data > 0) + return 0; + + usleep_range(500, 510); + } while (fifo_retry_count--); + + if (fifo_outstanding_data == 0) { + dev_err_ratelimited(swrm->dev, "%s err read underflow\n", __func__); + return -EIO; + } + + return 0; +} + +static int swrm_wait_for_wr_fifo_avail(struct qcom_swrm_ctrl *swrm) +{ + u32 fifo_outstanding_cmds, value; + int fifo_retry_count = SWR_OVERFLOW_RETRY_COUNT; + + do { + /* Check for fifo overflow during write */ + swrm->reg_read(swrm, SWRM_CMD_FIFO_STATUS, &value); + fifo_outstanding_cmds = FIELD_GET(SWRM_WR_CMD_FIFO_CNT_MASK, value); + + /* Check for space in write fifo before writing */ + if (fifo_outstanding_cmds < swrm->wr_fifo_depth) + return 0; + + usleep_range(500, 510); + } while (fifo_retry_count--); + + if (fifo_outstanding_cmds == swrm->wr_fifo_depth) { + dev_err_ratelimited(swrm->dev, "%s err write overflow\n", __func__); + return -EIO; + } + + return 0; +} static int qcom_swrm_cmd_fifo_wr_cmd(struct qcom_swrm_ctrl *swrm, u8 cmd_data, u8 dev_addr, u16 reg_addr) @@ -256,6 +312,9 @@ static int qcom_swrm_cmd_fifo_wr_cmd(struct qcom_swrm_ctrl *swrm, u8 cmd_data, dev_addr, reg_addr); } + if (swrm_wait_for_wr_fifo_avail(swrm)) + return SDW_CMD_FAIL_OTHER; + /* Its assumed that write is okay as we do not get any status back */ swrm->reg_write(swrm, SWRM_CMD_FIFO_WR_CMD, val); @@ -295,6 +354,9 @@ static int qcom_swrm_cmd_fifo_rd_cmd(struct qcom_swrm_ctrl *swrm, /* wait for FIFO RD CMD complete to avoid overflow */ usleep_range(250, 255); + if (swrm_wait_for_rd_fifo_avail(swrm)) + return SDW_CMD_FAIL_OTHER; + do { swrm->reg_read(swrm, SWRM_CMD_FIFO_RD_FIFO_ADDR, &cmd_data); rval[0] = cmd_data & 0xFF; @@ -586,6 +648,10 @@ static int qcom_swrm_init(struct qcom_swrm_ctrl *ctrl) SWRM_INTERRUPT_STATUS_RMSK); } ctrl->slave_status = 0; + ctrl->reg_read(ctrl, SWRM_COMP_PARAMS, &val); + ctrl->rd_fifo_depth = FIELD_GET(SWRM_COMP_PARAMS_RD_FIFO_DEPTH, val); + ctrl->wr_fifo_depth = FIELD_GET(SWRM_COMP_PARAMS_WR_FIFO_DEPTH, val); + return 0; } -- 2.21.0

3 3

[PATCH] soundwire: qcom: cleanup internal port config indexing
by Srinivas Kandagatla 06 Apr '21

06 Apr '21

Internally used portconfig array for storing port bandwidth params starts from offset zero. However port zero is not really used and we also copy the bus parameters to offset zero. So basically we endup with a code which has to subtract 1 from port number to get to port parameters. This is bit confusing to the reader so, make this bit more obvious by only copying the parameters to offset 1 instead of zero. This will avoid doing -1 every time when we try to get port params. Similar thing has been recently done with din/dout_port_mask. Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla(a)linaro.org> --- drivers/soundwire/qcom.c | 27 ++++++++++++++------------- 1 file changed, 14 insertions(+), 13 deletions(-) diff --git a/drivers/soundwire/qcom.c b/drivers/soundwire/qcom.c index 94fd58f9dda3..348d9a46f850 100644 --- a/drivers/soundwire/qcom.c +++ b/drivers/soundwire/qcom.c @@ -724,7 +724,7 @@ static int qcom_swrm_transport_params(struct sdw_bus *bus, int reg = SWRM_DP_PORT_CTRL_BANK((params->port_num), bank); int ret; - pcfg = &ctrl->pconfig[params->port_num - 1]; + pcfg = &ctrl->pconfig[params->port_num]; value = pcfg->off1 << SWRM_DP_PORT_CTRL_OFFSET1_SHFT; value |= pcfg->off2 << SWRM_DP_PORT_CTRL_OFFSET2_SHFT; @@ -811,11 +811,11 @@ static int qcom_swrm_compute_params(struct sdw_bus *bus) struct qcom_swrm_port_config *pcfg; struct sdw_slave *slave; unsigned int m_port; - int i = 0; + int i = 1; list_for_each_entry(m_rt, &bus->m_rt_list, bus_node) { list_for_each_entry(p_rt, &m_rt->port_list, port_node) { - pcfg = &ctrl->pconfig[p_rt->num - 1]; + pcfg = &ctrl->pconfig[p_rt->num]; p_rt->transport_params.port_num = p_rt->num; if (pcfg->word_length != SWR_INVALID_PARAM) { sdw_fill_port_params(&p_rt->port_params, @@ -832,7 +832,7 @@ static int qcom_swrm_compute_params(struct sdw_bus *bus) m_port = slave->m_port_map[p_rt->num]; /* port config starts at offset 0 so -1 from actual port number */ if (m_port) - pcfg = &ctrl->pconfig[m_port - 1]; + pcfg = &ctrl->pconfig[m_port]; else pcfg = &ctrl->pconfig[i]; p_rt->transport_params.port_num = p_rt->num; @@ -1167,15 +1167,16 @@ static int qcom_swrm_get_port_config(struct qcom_swrm_ctrl *ctrl) of_property_read_u8_array(np, "qcom,ports-lane-control", lane_control, nports); for (i = 0; i < nports; i++) { - ctrl->pconfig[i].si = si[i]; - ctrl->pconfig[i].off1 = off1[i]; - ctrl->pconfig[i].off2 = off2[i]; - ctrl->pconfig[i].bp_mode = bp_mode[i]; - ctrl->pconfig[i].hstart = hstart[i]; - ctrl->pconfig[i].hstop = hstop[i]; - ctrl->pconfig[i].word_length = word_length[i]; - ctrl->pconfig[i].blk_group_count = blk_group_count[i]; - ctrl->pconfig[i].lane_control = lane_control[i]; + /* Valid port number range is from 1-14 */ + ctrl->pconfig[i + 1].si = si[i]; + ctrl->pconfig[i + 1].off1 = off1[i]; + ctrl->pconfig[i + 1].off2 = off2[i]; + ctrl->pconfig[i + 1].bp_mode = bp_mode[i]; + ctrl->pconfig[i + 1].hstart = hstart[i]; + ctrl->pconfig[i + 1].hstop = hstop[i]; + ctrl->pconfig[i + 1].word_length = word_length[i]; + ctrl->pconfig[i + 1].blk_group_count = blk_group_count[i]; + ctrl->pconfig[i + 1].lane_control = lane_control[i]; } return 0; -- 2.21.0

2 1

[PATCH] ALSA: core: control_led: fix memory leak in snd_ctl_led_set_state()
by Takashi Sakamoto 06 Apr '21

06 Apr '21

When inquired control element is not in led group, snd_ctl_led_set_state() attempts to add the element into the group, according to function arguments. Although an memory object is allocated for led instance, it's left as is without being released. This commit fixes the memory leak. Fixes: 22d8de62f11b ("ALSA: control - add generic LED trigger module as the new control layer") Signed-off-by: Takashi Sakamoto <o-takashi(a)sakamocchi.jp> --- sound/core/control_led.c | 1 + 1 file changed, 1 insertion(+) diff --git a/sound/core/control_led.c b/sound/core/control_led.c index 788fd9e275e0..b97f118cd54e 100644 --- a/sound/core/control_led.c +++ b/sound/core/control_led.c @@ -161,6 +161,7 @@ static void snd_ctl_led_set_state(struct snd_card *card, unsigned int access, list_add(&lctl->list, &led->controls); UPDATE_ROUTE(route, snd_ctl_led_get(lctl)); } + kfree(lctl); } mutex_unlock(&snd_ctl_led_mutex); switch (led->mode) { -- 2.27.0

2 2

[PATCH v2 0/2] kbl_da7219_max9357a machine changes for wov and MST
by vamshi.krishna.gopal＠intel.com 05 Apr '21

05 Apr '21

From: Vamshi Krishna Gopal <vamshi.krishna.gopal(a)intel.corp-partner.google.com> Hi all, This patch series about creating dailink for Wake on voice functionality and also adding MST route changes. changes in v2 - Add MST route changes to kbl_da7219_max98927.c and kabl_rt5663_max98927.c - typo corrected in struct skylaye_refcap_ops to skylake_refcap_ops - Declare constraints_1600 struct and ch_mono variable v1 is here: https://patchwork.kernel.org/project/alsa-devel/cover/20210324175200.44922-… Mac Chiang (1): ASoc: Intel: board: add BE DAI link for WoV Vamshi Krishna Gopal (1): ASoC: Intel: kbl: Add MST route change to kbl machine drivers sound/soc/intel/boards/kbl_da7219_max98357a.c | 57 ++++++++++++++++++- sound/soc/intel/boards/kbl_da7219_max98927.c | 9 ++- sound/soc/intel/boards/kbl_rt5663_max98927.c | 4 ++ 3 files changed, 64 insertions(+), 6 deletions(-) -- 2.17.1

3 4

Front headphones jack not reporting state to Pulseaudio correctly (always unplugged)
by GitHub issues - opened 05 Apr '21

05 Apr '21

alsa-project/alsa-lib issue #130 was opened from jose1711: ALSA seems to always report state of front headphones jack as "unplugged" to PulseAudio (related bug report: https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/issues/1180). [alsa-info.txt](https://github.com/alsa-project/alsa-lib/files/6255379/alsa-… Issue URL : https://github.com/alsa-project/alsa-lib/issues/130 Repository URL: https://github.com/alsa-project/alsa-lib

1 0

[PATCH] Pioneer devices: engage implicit feedback sync for playback
by Geraldo 05 Apr '21

05 Apr '21

Hello everybody, This is a highly experimental patch on top of 5.12-rc4 that's supposed to engage implicit feedback sync on the playback for Pioneer devices. Without implicit feedback sync the inputs started glitching due to clock drift in about an hour in my Pioneer DJ DDJ-SR2. If you own a Pioneer device please test this patch and verify in the dyndbg logs if your capture(input) EP is being opened twice and if implicit_fb = 1 for the playback EP. I can't ask of you all to test this for hours and hours but if you wanted to have an excuse to digitize vinyl or spin up DVS this is your perfect one. Help open source! I'd like to thank Takashi Iwai for his tireless support in the Pioneer regression and beyond. Hopefully this highly experimental feature's code can be polished and mainlined eventually. --- implicit.c.git 2021-03-21 22:40:48.363417245 -0300 +++ implicit.c 2021-03-22 01:03:05.726729481 -0300 @@ -177,30 +177,31 @@ ifnum, alts); } -/* Playback and capture EPs on Pioneer devices share the same iface/altset, - * but they don't seem working with the implicit fb mode well, hence we - * just return as if the sync were already set up. +/* Pioneer devices: playback and capture streams sharing the same iface/altset */ -static int skip_pioneer_sync_ep(struct snd_usb_audio *chip, - struct audioformat *fmt, - struct usb_host_interface *alts) -{ - struct usb_endpoint_descriptor *epd; - - if (alts->desc.bNumEndpoints != 2) - return 0; - - epd = get_endpoint(alts, 1); - if (!usb_endpoint_is_isoc_in(epd) || - (epd->bmAttributes & USB_ENDPOINT_SYNCTYPE) != USB_ENDPOINT_SYNC_ASYNC || - ((epd->bmAttributes & USB_ENDPOINT_USAGE_MASK) != - USB_ENDPOINT_USAGE_DATA && - (epd->bmAttributes & USB_ENDPOINT_USAGE_MASK) != - USB_ENDPOINT_USAGE_IMPLICIT_FB)) - return 0; - return 1; /* don't handle with the implicit fb, just skip sync EP */ +static int add_pioneer_implicit_fb(struct snd_usb_audio *chip, + struct audioformat *fmt, + struct usb_host_interface *alts) +{ + struct usb_endpoint_descriptor *epd; + + if (alts->desc.bNumEndpoints != 2) + return 0; + + epd = get_endpoint(alts, 1); + + if (!usb_endpoint_is_isoc_in(epd) || + (epd->bmAttributes & USB_ENDPOINT_SYNCTYPE) != USB_ENDPOINT_SYNC_ASYNC || + ((epd->bmAttributes & USB_ENDPOINT_USAGE_MASK) != + USB_ENDPOINT_USAGE_DATA && + (epd->bmAttributes & USB_ENDPOINT_USAGE_MASK) != + USB_ENDPOINT_USAGE_IMPLICIT_FB)) + return 0; + return add_implicit_fb_sync_ep(chip, fmt, epd->bEndpointAddress, 1, + alts->desc.bInterfaceNumber, alts); } + static int __add_generic_implicit_fb(struct snd_usb_audio *chip, struct audioformat *fmt, int iface, int altset) @@ -306,7 +307,7 @@ alts->desc.bInterfaceClass == USB_CLASS_VENDOR_SPEC && (USB_ID_VENDOR(chip->usb_id) == 0x2b73 || /* Pioneer */ USB_ID_VENDOR(chip->usb_id) == 0x08e4 /* Pioneer */)) { - if (skip_pioneer_sync_ep(chip, fmt, alts)) + if (add_pioneer_implicit_fb(chip, fmt, alts)) return 1; } @@ -339,8 +340,20 @@ struct audioformat *fmt, struct usb_host_interface *alts) { - if (fmt->endpoint & USB_DIR_IN) - return audioformat_capture_quirk(chip, fmt, alts); + bool isPioneer; + + if (alts->desc.bInterfaceClass == USB_CLASS_VENDOR_SPEC && + (USB_ID_VENDOR(chip->usb_id) == 0x2b73 || /* Pioneer */ + USB_ID_VENDOR(chip->usb_id) == 0x08e4 /* Pioneer */)) + isPioneer = true; + + if (fmt->endpoint & USB_DIR_IN) { + if (isPioneer == true) + return 1; + else + return audioformat_capture_quirk(chip, fmt, alts); + } + else return audioformat_implicit_fb_quirk(chip, fmt, alts); } --- endpoint.c.git 2021-03-21 22:40:54.800084101 -0300 +++ endpoint.c 2021-03-21 22:45:41.543425855 -0300 @@ -688,7 +688,7 @@ goto unlock; } - if (!ep->opened) { + if (ep->opened < 2) { if (is_sync_ep) { ep->iface = fp->sync_iface; ep->altsetting = fp->sync_altsetting;

1 3

ALC1220 and presence of ucm2/ucm.conf breaks audio
by GitHub issues - opened 04 Apr '21

04 Apr '21

alsa-project/alsa-ucm-conf issue #86 was opened from devnull0000: I'd want to make a better report but I am not an ALSA expert, so the issue is that after update I lose any sound and I use ALC1220 codec on MSI MEG X570 GODLIKE motherboard. Deep googling gave me the report: https://bugs.archlinux.org/task/67058 So indeed deletion of the file helped immediately. The bug is super confusing because it looks like the device is present at alsa level but absent on Pulse Audio and it is not shown in GUI at all. I am at alsa-ucm-conf-1.2.4, linux-5.11.11, pulseaudio 13.0, gentoo. Issue URL : https://github.com/alsa-project/alsa-ucm-conf/issues/86 Repository URL: https://github.com/alsa-project/alsa-ucm-conf

1 0