Skip to contentSkip to conversation

ASoC: Intel: Skylake: Critical IPC fixes and flow adjustments

PATCH In review [email protected] 3 participants 14 messages Jun 13, 2019 – Jun 26, 2019

Subsystems (3)
lore
Share & tools
lore

[email protected]

author maintainer
#0 lore raw

cover ASoC: Intel: Skylake: Critical IPC fixes and flow adjustments

From: Cezary Rojewski <[email protected]>

This set of changes contains critical IPC fixes:
- ASoC: Intel: Skylake: Read HIPCT extension before clearing DONE bit
- ASoC: Intel: Fix race condition in IPC rx list

as well as updates to common mechanisms to use recommended flows, most notably:
- ASoC: Intel: Skylake: Use recommended SDxFMT programming sequence
- ASoC: Intel: Skylake: Fix incorrect capture position reporting

While flow adjustments fix issues with 100% reproduction rate, the most crucial changes address issues occurring rarely - usually only under stress conditions. These, however, can easily cause unexpected behavior on DSP level.

Amadeusz Sławiński (1):
ASoC: Intel: Skylake: Reset pipeline before its deletion

Cezary Rojewski (3):
ASoC: Intel: Skylake: Fix incorrect capture position reporting
ASoC: Intel: Skylake: Read HIPCT extension before clearing DONE bit
ASoC: Intel: Common: Fix NULL dereference in tx_wait_done

Gustaw Lewandowski (1):
ASoC: Intel: Fix race condition in IPC rx list

Kamil Lulko (1):
ASoC: Intel: Skylake: Strip T and L from TLV IPCs

Paweł Harłoziński (1):
ASoC: Intel: Skylake: Use recommended SDxFMT programming sequence

 include/sound/hda_codec.h              |  3 +++
 sound/pci/hda/hda_intel.c              |  3 ---
 sound/soc/intel/common/sst-ipc.c       |  2 +-
 sound/soc/intel/skylake/cnl-sst.c      |  2 +-
 sound/soc/intel/skylake/skl-messages.c | 23 ++++++++++++++++-------
 sound/soc/intel/skylake/skl-pcm.c      | 14 +++++++++++++-
 sound/soc/intel/skylake/skl-sst-ipc.c  |  4 ++--
 sound/soc/intel/skylake/skl-topology.c | 22 +++++++++-------------
 sound/soc/intel/skylake/skl.c          | 21 +++++++++++++++++++++
 sound/soc/intel/skylake/skl.h          |  1 +
 10 files changed, 67 insertions(+), 28 deletions(-)
Signature · 6 lines
2.17.1

_______________________________________________
Alsa-devel mailing list
[email protected]
https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
author maintainer
#1 lore raw

PATCH 1/7 ASoC: Intel: Skylake: Fix incorrect capture position reporting

From: Cezary Rojewski <[email protected]>

HW recommends to set DUM bit on device power up, so that DPIB write request occurs every frame regardless of whether DPIB has changed or not. This addresses incorrect position reporting for capture streams.

---
 sound/soc/intel/skylake/skl.c | 21 +++++++++++++++++++++
 sound/soc/intel/skylake/skl.h |  1 +
 2 files changed, 22 insertions(+)
sound/soc/intel/skylake/skl.c +21 -0
index f864f7b3df3a..7174b5da4099 100644
@@ -192,6 +192,25 @@ void skl_update_d0i3c(struct device *dev, bool enable)
 			snd_hdac_chip_readb(bus, VS_D0I3C));
 }
 
+/**
+ * skl_dum_set - set DUM bit in EM2 register
+ * @bus: HD-audio core bus
+ *
+ * Addresses incorrect position reporting for capture streams.
+ * Used on device power up.
+ */
+static void skl_dum_set(struct hdac_bus *bus)
+{
+	/* For the DUM bit to be set, CRST needs to be out of reset state */
+	if (!(snd_hdac_chip_readb(bus, GCTL) & AZX_GCTL_RESET)) {
+		skl_enable_miscbdcge(bus->dev, false);
+		snd_hdac_bus_exit_link_reset(bus);
+		skl_enable_miscbdcge(bus->dev, true);
+	}
+
+	snd_hdac_chip_updatel(bus, VS_EM2, AZX_VS_EM2_DUM, AZX_VS_EM2_DUM);
+}
+
 /* called from IRQ */
 static void skl_stream_update(struct hdac_bus *bus, struct hdac_stream *hstr)
 {
@@ -299,6 +318,7 @@ static int _skl_resume(struct hdac_bus *bus)
 	struct skl *skl = bus_to_skl(bus);
 
 	skl_init_pci(skl);
+	skl_dum_set(bus);
 	skl_init_chip(bus, true);
 
 	return skl_resume_dsp(skl);
@@ -956,6 +976,7 @@ static int skl_first_init(struct hdac_bus *bus)
 
 	/* initialize chip */
 	skl_init_pci(skl);
+	skl_dum_set(bus);
 
 	return skl_init_chip(bus, true);
 }
sound/soc/intel/skylake/skl.h +1 -0
index 85f8bb6687dc..b92a7f8fe675 100644
@@ -46,6 +46,7 @@
 #define DMA_TRANSMITION_START	2
 #define DMA_TRANSMITION_STOP	3
 
+#define AZX_VS_EM2_DUM			BIT(23)
 #define AZX_REG_VS_EM2_L1SEN		BIT(13)
 
 struct skl_dsp_resource {
Signature · 1 line
2.17.1
author maintainer
#2 lore raw

PATCH 2/7 ASoC: Intel: Skylake: Use recommended SDxFMT programming sequence

In reply to [email protected]

From: Paweł Harłoziński <[email protected]>

For BXT platforms, the recommended sequence to program the SDxFMT is to first couple the stream, write the format and decouple again. For all other platforms said sequence remains unchanged.

To prevent code duplication, IS_BXT (and consequently IS_CFL) macro is relocated to hda_codec.h file so it can be accessed by SKL driver.

---
 include/sound/hda_codec.h         |  3 +++
 sound/pci/hda/hda_intel.c         |  3 ---
 sound/soc/intel/skylake/skl-pcm.c | 14 +++++++++++++-
 3 files changed, 16 insertions(+), 4 deletions(-)
include/sound/hda_codec.h +3 -0
index cc7c8d42d4fd..ad46a082b00f 100644
@@ -31,6 +31,9 @@
 #include <sound/hda_verbs.h>
 #include <sound/hda_regmap.h>
 
+#define IS_BXT(pci) ((pci)->vendor == 0x8086 && (pci)->device == 0x5a98)
+#define IS_CFL(pci) ((pci)->vendor == 0x8086 && (pci)->device == 0xa348)
+
 /*
  * Structures
  */
sound/pci/hda/hda_intel.c +0 -3
index 0741eae23f10..07144bcc7059 100644
@@ -373,9 +373,6 @@ enum {
 					((pci)->device == 0x0d0c) || \
 					((pci)->device == 0x160c))
 
-#define IS_BXT(pci) ((pci)->vendor == 0x8086 && (pci)->device == 0x5a98)
-#define IS_CFL(pci) ((pci)->vendor == 0x8086 && (pci)->device == 0xa348)
-
 static char *driver_short_names[] = {
 	[AZX_DRIVER_ICH] = "HDA Intel",
 	[AZX_DRIVER_PCH] = "HDA Intel PCH",
sound/soc/intel/skylake/skl-pcm.c +13 -1
index edc5ecfc0b55..cd254823813a 100644
@@ -141,6 +141,7 @@ static void skl_set_suspend_active(struct snd_pcm_substream *substream,
 int skl_pcm_host_dma_prepare(struct device *dev, struct skl_pipe_params *params)
 {
 	struct hdac_bus *bus = dev_get_drvdata(dev);
+	struct skl *skl = bus_to_skl(bus);
 	unsigned int format_val;
 	struct hdac_stream *hstream;
 	struct hdac_ext_stream *stream;
@@ -165,7 +166,18 @@ int skl_pcm_host_dma_prepare(struct device *dev, struct skl_pipe_params *params)
 	if (err < 0)
 		return err;
 
-	err = snd_hdac_stream_setup(hdac_stream(stream));
+	/*
+	 * The recommended SDxFMT programming sequence for BXT
+	 * platforms is to couple the stream before writing the format
+	 */
+	if (IS_BXT(skl->pci)) {
+		snd_hdac_ext_stream_decouple(bus, stream, false);
+		err = snd_hdac_stream_setup(hdac_stream(stream));
+		snd_hdac_ext_stream_decouple(bus, stream, true);
+	} else {
+		err = snd_hdac_stream_setup(hdac_stream(stream));
+	}
+
 	if (err < 0)
 		return err;
 
Signature · 6 lines
2.17.1

_______________________________________________
Alsa-devel mailing list
[email protected]
https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
author maintainer
#3 lore raw

PATCH 3/7 ASoC: Intel: Skylake: Read HIPCT extension before clearing DONE bit

In reply to [email protected]

From: Cezary Rojewski <[email protected]>

Host clears DONE bit to signal IPC target it has completed the operation. Once this is done, IPC target i.e. DSP may proceed with the next reply, filling registers with new portion of data.

Because of this, host should always read all registers prior to clearing DONE and BUSY bits to ensure no desynchronization happens the time in between clearing bits and reading message data (here, extension).

---
 sound/soc/intel/skylake/cnl-sst.c     | 2 +-
 sound/soc/intel/skylake/skl-sst-ipc.c | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
sound/soc/intel/skylake/cnl-sst.c +1 -1
index 245df1067ba8..759ea59615ca 100644
@@ -313,6 +313,7 @@ static irqreturn_t cnl_dsp_irq_thread_handler(int irq, void *context)
 
 	hipcida = sst_dsp_shim_read_unlocked(dsp, CNL_ADSP_REG_HIPCIDA);
 	hipctdr = sst_dsp_shim_read_unlocked(dsp, CNL_ADSP_REG_HIPCTDR);
+	hipctdd = sst_dsp_shim_read_unlocked(dsp, CNL_ADSP_REG_HIPCTDD);
 
 	/* reply message from dsp */
 	if (hipcida & CNL_ADSP_REG_HIPCIDA_DONE) {
@@ -332,7 +333,6 @@ static irqreturn_t cnl_dsp_irq_thread_handler(int irq, void *context)
 
 	/* new message from dsp */
 	if (hipctdr & CNL_ADSP_REG_HIPCTDR_BUSY) {
-		hipctdd = sst_dsp_shim_read_unlocked(dsp, CNL_ADSP_REG_HIPCTDD);
 		header.primary = hipctdr;
 		header.extension = hipctdd;
 		dev_dbg(dsp->dev, "IPC irq: Firmware respond primary:%x",
sound/soc/intel/skylake/skl-sst-ipc.c +1 -1
index 9f3ce73593ae..5c9206dc7932 100644
@@ -511,6 +511,7 @@ irqreturn_t skl_dsp_irq_thread_handler(int irq, void *context)
 
 	hipcie = sst_dsp_shim_read_unlocked(dsp, SKL_ADSP_REG_HIPCIE);
 	hipct = sst_dsp_shim_read_unlocked(dsp, SKL_ADSP_REG_HIPCT);
+	hipcte = sst_dsp_shim_read_unlocked(dsp, SKL_ADSP_REG_HIPCTE);
 
 	/* reply message from DSP */
 	if (hipcie & SKL_ADSP_REG_HIPCIE_DONE) {
@@ -530,7 +531,6 @@ irqreturn_t skl_dsp_irq_thread_handler(int irq, void *context)
 
 	/* New message from DSP */
 	if (hipct & SKL_ADSP_REG_HIPCT_BUSY) {
-		hipcte = sst_dsp_shim_read_unlocked(dsp, SKL_ADSP_REG_HIPCTE);
 		header.primary = hipct;
 		header.extension = hipcte;
 		dev_dbg(dsp->dev, "IPC irq: Firmware respond primary:%x\n",
Signature · 1 line
2.17.1
maintainer
#4 lore raw

Applied "ASoC: Intel: Skylake: Read HIPCT extension before clearing DONE bit" to the asoc tree

The patch

ASoC: Intel: Skylake: Read HIPCT extension before clearing DONE bit

has been applied to the asoc tree at

https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound.git for-5.3

All being well this means that it will be integrated into the linux-next tree (usually sometime in the next 24 hours) and sent to Linus during the next merge window (or sooner if it is a bug fix), however if problems are discovered then the patch may be dropped or reverted.

You may get further e-mails resulting from automated or manual testing and review of the tree, please engage with people reporting problems and send followup patches addressing any issues that are reported if needed.

If any updates are required or you are submitting further changes they should be sent as incremental updates against current git, existing patches will not be replaced.

Please add any relevant lists and maintainers to the CCs when replying to this mail.

Thanks,
Mark

Quoted 1 line

From 078759399ff74e2e6f5e208c61924d1b7d66e5d8 Mon Sep 17 00:00:00 2001

From: Cezary Rojewski <[email protected]>
Date: Thu, 13 Jun 2019 21:04:32 +0200
Subject: [PATCH] ASoC: Intel: Skylake: Read HIPCT extension before clearing
DONE bit

Host clears DONE bit to signal IPC target it has completed the operation. Once this is done, IPC target i.e. DSP may proceed with the next reply, filling registers with new portion of data.

Because of this, host should always read all registers prior to clearing DONE and BUSY bits to ensure no desynchronization happens the time in between clearing bits and reading message data (here, extension).

---
 sound/soc/intel/skylake/cnl-sst.c     | 2 +-
 sound/soc/intel/skylake/skl-sst-ipc.c | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
sound/soc/intel/skylake/cnl-sst.c +1 -1
index 245df1067ba8..759ea59615ca 100644
@@ -313,6 +313,7 @@ static irqreturn_t cnl_dsp_irq_thread_handler(int irq, void *context)
 
 	hipcida = sst_dsp_shim_read_unlocked(dsp, CNL_ADSP_REG_HIPCIDA);
 	hipctdr = sst_dsp_shim_read_unlocked(dsp, CNL_ADSP_REG_HIPCTDR);
+	hipctdd = sst_dsp_shim_read_unlocked(dsp, CNL_ADSP_REG_HIPCTDD);
 
 	/* reply message from dsp */
 	if (hipcida & CNL_ADSP_REG_HIPCIDA_DONE) {
@@ -332,7 +333,6 @@ static irqreturn_t cnl_dsp_irq_thread_handler(int irq, void *context)
 
 	/* new message from dsp */
 	if (hipctdr & CNL_ADSP_REG_HIPCTDR_BUSY) {
-		hipctdd = sst_dsp_shim_read_unlocked(dsp, CNL_ADSP_REG_HIPCTDD);
 		header.primary = hipctdr;
 		header.extension = hipctdd;
 		dev_dbg(dsp->dev, "IPC irq: Firmware respond primary:%x",
sound/soc/intel/skylake/skl-sst-ipc.c +1 -1
index 9f3ce73593ae..5c9206dc7932 100644
@@ -511,6 +511,7 @@ irqreturn_t skl_dsp_irq_thread_handler(int irq, void *context)
 
 	hipcie = sst_dsp_shim_read_unlocked(dsp, SKL_ADSP_REG_HIPCIE);
 	hipct = sst_dsp_shim_read_unlocked(dsp, SKL_ADSP_REG_HIPCT);
+	hipcte = sst_dsp_shim_read_unlocked(dsp, SKL_ADSP_REG_HIPCTE);
 
 	/* reply message from DSP */
 	if (hipcie & SKL_ADSP_REG_HIPCIE_DONE) {
@@ -530,7 +531,6 @@ irqreturn_t skl_dsp_irq_thread_handler(int irq, void *context)
 
 	/* New message from DSP */
 	if (hipct & SKL_ADSP_REG_HIPCT_BUSY) {
-		hipcte = sst_dsp_shim_read_unlocked(dsp, SKL_ADSP_REG_HIPCTE);
 		header.primary = hipct;
 		header.extension = hipcte;
 		dev_dbg(dsp->dev, "IPC irq: Firmware respond primary:%x\n",
Signature · 1 line
2.20.1
author maintainer
#5 lore raw

PATCH 4/7 ASoC: Intel: Fix race condition in IPC rx list

In reply to [email protected]

From: Gustaw Lewandowski <[email protected]>

Since there are multiple IPCs being sent in a short span of time, there is a possibility of more than one message being on the Rx list after receiving response from firmware. In such cases, when the first notification of interrupt from firmware is received, driver retrieves the message from the Rx list but does not delete it from the list till the next lock. In the meantime, when another interrupt is received from the firmware, driver is reading the previous message again since the previous message has not been removed from the list.

---
 sound/soc/intel/skylake/skl-sst-ipc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
sound/soc/intel/skylake/skl-sst-ipc.c +1 -1
index 5c9206dc7932..5094205a243f 100644
@@ -344,6 +344,7 @@ static struct ipc_message *skl_ipc_reply_get_msg(struct sst_generic_ipc *ipc,
 
 	msg = list_first_entry(&ipc->rx_list, struct ipc_message, list);
 
+	list_del(&msg->list);
 out:
 	return msg;
 
@@ -488,7 +489,6 @@ void skl_ipc_process_reply(struct sst_generic_ipc *ipc,
 	}
 
 	spin_lock_irqsave(&ipc->dsp->spinlock, flags);
-	list_del(&msg->list);
 	sst_ipc_tx_msg_reply_complete(ipc, msg);
 	spin_unlock_irqrestore(&ipc->dsp->spinlock, flags);
 }
Signature · 1 line
2.17.1
maintainer
#6 lore raw

Applied "ASoC: Intel: Fix race condition in IPC rx list" to the asoc tree

The patch

ASoC: Intel: Fix race condition in IPC rx list

has been applied to the asoc tree at

https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound.git for-5.3

All being well this means that it will be integrated into the linux-next tree (usually sometime in the next 24 hours) and sent to Linus during the next merge window (or sooner if it is a bug fix), however if problems are discovered then the patch may be dropped or reverted.

You may get further e-mails resulting from automated or manual testing and review of the tree, please engage with people reporting problems and send followup patches addressing any issues that are reported if needed.

If any updates are required or you are submitting further changes they should be sent as incremental updates against current git, existing patches will not be replaced.

Please add any relevant lists and maintainers to the CCs when replying to this mail.

Thanks,
Mark

Quoted 1 line

From 26ae20490809db30677dfd54f81a73ce77ba2df1 Mon Sep 17 00:00:00 2001

From: Gustaw Lewandowski <[email protected]>
Date: Thu, 13 Jun 2019 21:04:33 +0200
Subject: [PATCH] ASoC: Intel: Fix race condition in IPC rx list

Since there are multiple IPCs being sent in a short span of time, there is a possibility of more than one message being on the Rx list after receiving response from firmware. In such cases, when the first notification of interrupt from firmware is received, driver retrieves the message from the Rx list but does not delete it from the list till the next lock. In the meantime, when another interrupt is received from the firmware, driver is reading the previous message again since the previous message has not been removed from the list.

---
 sound/soc/intel/skylake/skl-sst-ipc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
sound/soc/intel/skylake/skl-sst-ipc.c +1 -1
index 5c9206dc7932..5094205a243f 100644
@@ -344,6 +344,7 @@ static struct ipc_message *skl_ipc_reply_get_msg(struct sst_generic_ipc *ipc,
 
 	msg = list_first_entry(&ipc->rx_list, struct ipc_message, list);
 
+	list_del(&msg->list);
 out:
 	return msg;
 
@@ -488,7 +489,6 @@ void skl_ipc_process_reply(struct sst_generic_ipc *ipc,
 	}
 
 	spin_lock_irqsave(&ipc->dsp->spinlock, flags);
-	list_del(&msg->list);
 	sst_ipc_tx_msg_reply_complete(ipc, msg);
 	spin_unlock_irqrestore(&ipc->dsp->spinlock, flags);
 }
Signature · 1 line
2.20.1
author maintainer
#7 lore raw

PATCH 5/7 ASoC: Intel: Common: Fix NULL dereference in tx_wait_done

In reply to [email protected]

From: Cezary Rojewski <[email protected]>

rx_data and rx_bytes present for tx_wait_done are optional parameters. If not provided, function should not attempt to copy received data. This change fixes memcpy NULL pointer dereference issue occurring when optional rx_data is NULL while received message size is non-zero.

---
 sound/soc/intel/common/sst-ipc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
sound/soc/intel/common/sst-ipc.c +1 -1
index dcff13802c00..fc3340f1da3a 100644
@@ -71,7 +71,7 @@ static int tx_wait_done(struct sst_generic_ipc *ipc,
 	} else {
 
 		/* copy the data returned from DSP */
-		if (msg->rx_size)
+		if (rx_data)
 			memcpy(rx_data, msg->rx_data, msg->rx_size);
 		ret = msg->errno;
 	}
Signature · 1 line
2.17.1
maintainer
#8 lore raw

Applied "ASoC: Intel: Common: Fix NULL dereference in tx_wait_done" to the asoc tree

The patch

ASoC: Intel: Common: Fix NULL dereference in tx_wait_done

has been applied to the asoc tree at

https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound.git for-5.3

All being well this means that it will be integrated into the linux-next tree (usually sometime in the next 24 hours) and sent to Linus during the next merge window (or sooner if it is a bug fix), however if problems are discovered then the patch may be dropped or reverted.

You may get further e-mails resulting from automated or manual testing and review of the tree, please engage with people reporting problems and send followup patches addressing any issues that are reported if needed.

If any updates are required or you are submitting further changes they should be sent as incremental updates against current git, existing patches will not be replaced.

Please add any relevant lists and maintainers to the CCs when replying to this mail.

Thanks,
Mark

Quoted 1 line

From 9f4f42d710d417745cff05845f93370126f77ff7 Mon Sep 17 00:00:00 2001

From: Cezary Rojewski <[email protected]>
Date: Thu, 13 Jun 2019 21:04:34 +0200
Subject: [PATCH] ASoC: Intel: Common: Fix NULL dereference in tx_wait_done

rx_data and rx_bytes present for tx_wait_done are optional parameters. If not provided, function should not attempt to copy received data. This change fixes memcpy NULL pointer dereference issue occurring when optional rx_data is NULL while received message size is non-zero.

---
 sound/soc/intel/common/sst-ipc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
sound/soc/intel/common/sst-ipc.c +1 -1
index dcff13802c00..fc3340f1da3a 100644
@@ -71,7 +71,7 @@ static int tx_wait_done(struct sst_generic_ipc *ipc,
 	} else {
 
 		/* copy the data returned from DSP */
-		if (msg->rx_size)
+		if (rx_data)
 			memcpy(rx_data, msg->rx_data, msg->rx_size);
 		ret = msg->errno;
 	}
Signature · 1 line
2.20.1
author maintainer
#9 lore raw

PATCH 6/7 ASoC: Intel: Skylake: Reset pipeline before its deletion

In reply to [email protected]

From: Amadeusz Sławiński <[email protected]>

Before actual deletion, pipeline should enter RESET state. Currently, pipe skips this checkpoint and goes straight to the finish line. This is not the expected path by the FW, so correct it.

---
 sound/soc/intel/skylake/skl-messages.c | 23 ++++++++++++++++-------
 1 file changed, 16 insertions(+), 7 deletions(-)
sound/soc/intel/skylake/skl-messages.c +16 -7
index df01dc952521..79baf90e6116 100644
@@ -1265,10 +1265,10 @@ int skl_create_pipeline(struct skl_sst *ctx, struct skl_pipe *pipe)
 }
 
 /*
- * A pipeline needs to be deleted on cleanup. If a pipeline is running, then
- * pause the pipeline first and then delete it
- * The pipe delete is done by sending delete pipeline IPC. DSP will stop the
- * DMA engines and releases resources
+ * A pipeline needs to be deleted on cleanup. If a pipeline is running,
+ * then pause it first. Before actual deletion, pipeline should enter
+ * reset state. Finish the procedure by sending delete pipeline IPC.
+ * DSP will stop the DMA engines and release resources
  */
 int skl_delete_pipe(struct skl_sst *ctx, struct skl_pipe *pipe)
 {
@@ -1276,6 +1276,10 @@ int skl_delete_pipe(struct skl_sst *ctx, struct skl_pipe *pipe)
 
 	dev_dbg(ctx->dev, "%s: pipe = %d\n", __func__, pipe->ppl_id);
 
+	/* If pipe was not created in FW, do not try to delete it */
+	if (pipe->state < SKL_PIPE_CREATED)
+		return 0;
+
 	/* If pipe is started, do stop the pipe in FW. */
 	if (pipe->state >= SKL_PIPE_STARTED) {
 		ret = skl_set_pipe_state(ctx, pipe, PPL_PAUSED);
@@ -1287,9 +1291,14 @@ int skl_delete_pipe(struct skl_sst *ctx, struct skl_pipe *pipe)
 		pipe->state = SKL_PIPE_PAUSED;
 	}
 
-	/* If pipe was not created in FW, do not try to delete it */
-	if (pipe->state < SKL_PIPE_CREATED)
-		return 0;
+	/* reset pipe state before deletion */
+	ret = skl_set_pipe_state(ctx, pipe, PPL_RESET);
+	if (ret < 0) {
+		dev_err(ctx->dev, "Failed to reset pipe ret=%d\n", ret);
+		return ret;
+	}
+
+	pipe->state = SKL_PIPE_RESET;
 
 	ret = skl_ipc_delete_pipeline(&ctx->ipc, pipe->ppl_id);
 	if (ret < 0) {
Signature · 6 lines
2.17.1

_______________________________________________
Alsa-devel mailing list
[email protected]
https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
author maintainer
#10 lore raw

PATCH 7/7 ASoC: Intel: Skylake: Strip T and L from TLV IPCs

In reply to [email protected]

From: Kamil Lulko <[email protected]>

cAVS modules do not require Type and Length header within the set_module_params IPC. This is also true for Vendor modules. The userspace (like tinymix) always appends this header to TLV controls which are used for set_module_params. Simply assume this header is always present in the payload and omit it from the IPC.

---
 sound/soc/intel/skylake/skl-topology.c | 22 +++++++++-------------
 1 file changed, 9 insertions(+), 13 deletions(-)
sound/soc/intel/skylake/skl-topology.c +9 -13
index c69d999d7bf1..76e971b8536a 100644
@@ -1492,22 +1492,18 @@ static int skl_tplg_tlv_control_set(struct snd_kcontrol *kcontrol,
 	struct skl *skl = get_skl_ctx(w->dapm->dev);
 
 	if (ac->params) {
+		/*
+		 * Widget data is expected to be stripped of T and L
+		 */
+		size -= 2 * sizeof(unsigned int);
+		data += 2;
+
 		if (size > ac->max)
 			return -EINVAL;
-
 		ac->size = size;
-		/*
-		 * if the param_is is of type Vendor, firmware expects actual
-		 * parameter id and size from the control.
-		 */
-		if (ac->param_id == SKL_PARAM_VENDOR_ID) {
-			if (copy_from_user(ac->params, data, size))
-				return -EFAULT;
-		} else {
-			if (copy_from_user(ac->params,
-					   data + 2, size))
-				return -EFAULT;
-		}
+
+		if (copy_from_user(ac->params, data, size))
+			return -EFAULT;
 
 		if (w->power)
 			return skl_set_module_params(skl->skl_sst,
Signature · 1 line
2.17.1
maintainer
#11 lore raw

Applied "ASoC: Intel: Skylake: Strip T and L from TLV IPCs" to the asoc tree

The patch

ASoC: Intel: Skylake: Strip T and L from TLV IPCs

has been applied to the asoc tree at

https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound.git for-5.3

All being well this means that it will be integrated into the linux-next tree (usually sometime in the next 24 hours) and sent to Linus during the next merge window (or sooner if it is a bug fix), however if problems are discovered then the patch may be dropped or reverted.

You may get further e-mails resulting from automated or manual testing and review of the tree, please engage with people reporting problems and send followup patches addressing any issues that are reported if needed.

If any updates are required or you are submitting further changes they should be sent as incremental updates against current git, existing patches will not be replaced.

Please add any relevant lists and maintainers to the CCs when replying to this mail.

Thanks,
Mark

Quoted 1 line

From a8cd7066f0422f378902770034ddac1720d0e032 Mon Sep 17 00:00:00 2001

From: Kamil Lulko <[email protected]>
Date: Thu, 13 Jun 2019 21:04:36 +0200
Subject: [PATCH] ASoC: Intel: Skylake: Strip T and L from TLV IPCs

cAVS modules do not require Type and Length header within the set_module_params IPC. This is also true for Vendor modules. The userspace (like tinymix) always appends this header to TLV controls which are used for set_module_params. Simply assume this header is always present in the payload and omit it from the IPC.

---
 sound/soc/intel/skylake/skl-topology.c | 22 +++++++++-------------
 1 file changed, 9 insertions(+), 13 deletions(-)
sound/soc/intel/skylake/skl-topology.c +9 -13
index 99825dda34af..c353eb14ce36 100644
@@ -1492,22 +1492,18 @@ static int skl_tplg_tlv_control_set(struct snd_kcontrol *kcontrol,
 	struct skl *skl = get_skl_ctx(w->dapm->dev);
 
 	if (ac->params) {
+		/*
+		 * Widget data is expected to be stripped of T and L
+		 */
+		size -= 2 * sizeof(unsigned int);
+		data += 2;
+
 		if (size > ac->max)
 			return -EINVAL;
-
 		ac->size = size;
-		/*
-		 * if the param_is is of type Vendor, firmware expects actual
-		 * parameter id and size from the control.
-		 */
-		if (ac->param_id == SKL_PARAM_VENDOR_ID) {
-			if (copy_from_user(ac->params, data, size))
-				return -EFAULT;
-		} else {
-			if (copy_from_user(ac->params,
-					   data + 2, size))
-				return -EFAULT;
-		}
+
+		if (copy_from_user(ac->params, data, size))
+			return -EFAULT;
 
 		if (w->power)
 			return skl_set_module_params(skl->skl_sst,
Signature · 1 line
2.20.1
#13 lore raw
Quoted 8 lines · Mark Brown
Quoted 5 lines · On Thu, Jun 13, 2019 at 09:04:29PM +0200, [email protected] wrote:

From: Cezary Rojewski <[email protected]>

This set of changes contains critical IPC fixes:
- ASoC: Intel: Skylake: Read HIPCT extension before clearing DONE bit
- ASoC: Intel: Fix race condition in IPC rx list

Pierre, any thoughts on this series?

This series I looked at in detail and it was the basis for the recent SOF cleanup on IPC. There are still some improvements being debated on when exactly the IPC interrupts need to be unmasked - Guennadi suggested the interrupt need to be re-enabled at the beginning of the interrupt thread to avoid missing events - but this can come as bullet-proof hardening patches later for both SOF and Skylake. For now this is a good first pass, feel free to take the following tag for the series