ASoC: Intel: avs: Fixes and cleanups
Subsystems (3)
Cezary Rojewski
maintainercover ASoC: Intel: avs: Fixes and cleanups
A set of loosely connected changes, fixing few outstanding issues as well as improving readability of the existing code.
The fixes lead the series, first five patches. The goal is to make sure proper read() is used when accessing the registers, probe() and remove() sequences for HDAudio streaming are synced, minimal AudioDSP firmware version points to correct values and recent additions to the topology are parsed properly.
The only patch that points to 'new functionality' is:
ASoC: Intel: avs: Update ASRC definition
as with the struct definition updates, one can utilize the ASRC module in both streaming directions now (previously limited to Capture). Everything else either improves the logging or provides comments vital for long-term maintenance of the code.
Amadeusz Sławiński (4):
ASoC: Intel: avs: Fix init-config parsing
ASoC: Intel: avs: Print IPC error messages in lower layer
ASoC: Intel: avs: Add MODULE_FIRMWARE to inform about FW
ASoC: Intel: avs: Update ASRC definition
Cezary Rojewski (9):
ASoC: Intel: avs: Do not readq() u32 registers
ASoC: Intel: avs: Fix the minimum firmware version numbers
ASoC: Intel: avs: Fix theoretical infinite loop
ASoC: Intel: avs: Update hda component teardown sequences
ASoC: Intel: avs: Clearly state assumptions of hw_params()
ASoC: Intel: avs: Improve logging of firmware loading
ASoC: Intel: avs: Adjust DSP status register names
ASoC: Intel: avs: Adjust IPC traces
ASoC: Intel: avs: Add missing includes
sound/soc/intel/avs/apl.c | 2 +- sound/soc/intel/avs/core.c | 24 +++++++++++++++------ sound/soc/intel/avs/debugfs.c | 1 + sound/soc/intel/avs/ipc.c | 25 +++++++++++++--------- sound/soc/intel/avs/loader.c | 36 ++++++++++++++++--------------- sound/soc/intel/avs/messages.c | 22 +++++++++++++------ sound/soc/intel/avs/messages.h | 3 +-- sound/soc/intel/avs/pcm.c | 5 ++++- sound/soc/intel/avs/registers.h | 2 +- sound/soc/intel/avs/topology.c | 4 ++-- sound/soc/intel/avs/trace.h | 38 +++++++++++++++++---------------- 11 files changed, 97 insertions(+), 65 deletions(-)
Signature · 1 line
2.25.1
Cezary Rojewski
maintainerPATCH 1/13 ASoC: Intel: avs: Do not readq() u32 registers
Register reporting ROM status is 4-bytes wide.
--- sound/soc/intel/avs/loader.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
sound/soc/intel/avs/loader.c +1 -1
@@ -308,7 +308,7 @@ avs_hda_init_rom(struct avs_dev *adev, unsigned int dma_id, bool purge) }/* await ROM init */ - ret =snd_hdac_adsp_readq_poll (adev, spec->sram->rom_status_offset, reg, + ret =snd_hdac_adsp_readl_poll (adev, spec->sram->rom_status_offset, reg, (reg &0xF ) ==AVS_ROM_INIT_DONE || (reg &0xF ) ==APL_ROM_FW_ENTERED ,AVS_ROM_INIT_POLLING_US ,APL_ROM_INIT_TIMEOUT_US );
Signature · 1 line
2.25.1
Cezary Rojewski
maintainerPATCH 2/13 ASoC: Intel: avs: Fix the minimum firmware version numbers
In reply to Cezary RojewskiFor few TGL-based platforms the minor version number for AudioDSP firmware is incorrect forcing users to utilize ignore_fw_version module parameter.
--- sound/soc/intel/avs/core.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-)
sound/soc/intel/avs/core.c +7 -7
@@ -829,10 +829,10 @@ static const struct avs_spec jsl_desc = { .hipc = &cnl_hipc_spec, }; -#define AVS_TGL_BASED_SPEC (sname) \ +#define AVS_TGL_BASED_SPEC (sname, min) \ static const struct avs_spec sname##_desc = { \ .name = #sname, \ - .min_fw_version = { 10, 29, 0, 5646 }, \ + .min_fw_version = { 10, min, 0, 5646 }, \ .dsp_ops = &avs_tgl_dsp_ops, \ .core_init_mask = 1, \ .attributes = AVS_PLATATTR_IMR, \
@@ -840,11 +840,11 @@ static const struct avs_spec sname##_desc = { \ .hipc = &cnl_hipc_spec, \ } -AVS_TGL_BASED_SPEC (lkf); -AVS_TGL_BASED_SPEC (tgl); -AVS_TGL_BASED_SPEC (ehl); -AVS_TGL_BASED_SPEC (adl); -AVS_TGL_BASED_SPEC (adl_n); +AVS_TGL_BASED_SPEC (lkf,28 ); +AVS_TGL_BASED_SPEC (tgl,29 ); +AVS_TGL_BASED_SPEC (ehl,30 ); +AVS_TGL_BASED_SPEC (adl,35 ); +AVS_TGL_BASED_SPEC (adl_n,35 );static const struct pci_device_id avs_ids[] = { {PCI_DEVICE_DATA (INTEL ,HDA_SKL_LP , &skl_desc) },
Signature · 1 line
2.25.1
Cezary Rojewski
maintainerPATCH 3/13 ASoC: Intel: avs: Fix theoretical infinite loop
In reply to Cezary RojewskiWhile 'stack_dump_size' is a u32 bitfield of 16 bits, u32 has a bigger upper bound than the type u16 of loop counter 'offset' what in theory may lead to infinite loop condition.
Found out by Coverity static analyzer.
--- sound/soc/intel/avs/apl.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
sound/soc/intel/avs/apl.c +1 -1
@@ -125,7 +125,7 @@ int avs_apl_coredump(struct avs_dev *adev, union avs_notify_msg *msg)struct avs_apl_log_buffer_layout layout;void __iomem *addr, *buf;size_t dump_size; -u16 offset =0 ; +u32 offset =0 ;u8 *dump, *pos; dump_size =AVS_FW_REGS_SIZE + msg->ext.coredump.stack_dump_size;
Signature · 1 line
2.25.1
Cezary Rojewski
maintainerPATCH 4/13 ASoC: Intel: avs: Fix init-config parsing
In reply to Cezary RojewskiFrom: Amadeusz Sławiński <[email protected]>
When parsing init configs correct token should be looked up.
--- sound/soc/intel/avs/topology.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
sound/soc/intel/avs/topology.c +2 -2
@@ -1466,7 +1466,7 @@ avs_tplg_path_template_create(struct snd_soc_component *comp, struct avs_tplg *ostatic const struct avs_tplg_token_parser mod_init_config_parsers[] = { { - .token =AVS_TKN_MOD_INIT_CONFIG_ID_U32 , + .token =AVS_TKN_INIT_CONFIG_ID_U32 , .type =SND_SOC_TPLG_TUPLE_TYPE_WORD , .offset = offsetof(struct avs_tplg_init_config , id), .parse = avs_parse_word_token,
@@ -1519,7 +1519,7 @@ static int avs_tplg_parse_initial_configs(struct snd_soc_component *comp, esize =le32_to_cpu (tuples->size) +le32_to_cpu (tmp->size); ret =parse_dictionary_entries (comp, tuples, esize, config,1 ,sizeof (*config), -AVS_TKN_MOD_INIT_CONFIG_ID_U32 , +AVS_TKN_INIT_CONFIG_ID_U32 , mod_init_config_parsers,ARRAY_SIZE (mod_init_config_parsers));
Signature · 1 line
2.25.1
Cezary Rojewski
maintainerPATCH 5/13 ASoC: Intel: avs: Update hda component teardown sequences
In reply to Cezary RojewskiIf case of failure cleanup recently created DAI and while at it, adjust the remove() operation to match operation order of the probe() function.
--- sound/soc/intel/avs/pcm.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
sound/soc/intel/avs/pcm.c +2 -1
@@ -1564,6 +1564,7 @@ static int avs_component_hda_probe(struct snd_soc_component *component)if (ret <0 ) {dev_err (component->dev,"create widgets failed: %d\n" , ret); +snd_soc_unregister_dai (dai); goto exit; } }
@@ -1578,8 +1579,8 @@ static int avs_component_hda_probe(struct snd_soc_component *component)static void avs_component_hda_remove (struct snd_soc_component *component) { -avs_component_hda_unregister_dais (component);avs_component_remove (component); +avs_component_hda_unregister_dais (component); } static int avs_component_hda_open(struct snd_soc_component *component,
Signature · 1 line
2.25.1
Cezary Rojewski
maintainerPATCH 6/13 ASoC: Intel: avs: Print IPC error messages in lower layer
In reply to Cezary RojewskiFrom: Amadeusz Sławiński <[email protected]>
It is preferred to send error message in handler itself instead of leaving it to caller.
--- sound/soc/intel/avs/loader.c | 8 ++------ sound/soc/intel/avs/messages.c | 22 ++++++++++++++++------ 2 files changed, 18 insertions(+), 12 deletions(-)
sound/soc/intel/avs/loader.c +2 -6
@@ -675,16 +675,12 @@ int avs_dsp_first_boot_firmware(struct avs_dev *adev) } ret =avs_ipc_get_hw_config (adev, &adev->hw_cfg); -if (ret) { -dev_err (adev->dev,"get hw cfg failed: %d\n" , ret); +if (ret)return AVS_IPC_RET (ret); - } ret =avs_ipc_get_fw_config (adev, &adev->fw_cfg); -if (ret) { -dev_err (adev->dev,"get fw cfg failed: %d\n" , ret); +if (ret)return AVS_IPC_RET (ret); - } adev->core_refs =devm_kcalloc (adev->dev, adev->hw_cfg.dsp_cores,sizeof (*adev->core_refs),GFP_KERNEL );
sound/soc/intel/avs/messages.c +16 -6
@@ -400,10 +400,12 @@ int avs_ipc_get_fw_config(struct avs_dev *adev, struct avs_fw_cfg *cfg)AVS_BASEFW_FIRMWARE_CONFIG ,NULL ,0 , &payload, &payload_size);if (ret) -return ret; + goto err;/* Non-zero payload expected for FIRMWARE_CONFIG. */ -if (!payload_size) -return -EREMOTEIO ; +if (!payload_size) { + ret = -EREMOTEIO ; + goto err; + }while (offset < payload_size) { tlv = (struct avs_tlv *)(payload + offset);
@@ -502,6 +504,9 @@ int avs_ipc_get_fw_config(struct avs_dev *adev, struct avs_fw_cfg *cfg)/* No longer needed, free it as it's owned by the get_large_config() caller. */ kfree (payload); +err: +if (ret) +dev_err (adev->dev,"get fw cfg failed: %d\n" , ret);return ret; }
@@ -517,10 +522,12 @@ int avs_ipc_get_hw_config(struct avs_dev *adev, struct avs_hw_cfg *cfg)AVS_BASEFW_HARDWARE_CONFIG ,NULL ,0 , &payload, &payload_size);if (ret) -return ret; + goto err;/* Non-zero payload expected for HARDWARE_CONFIG. */ -if (!payload_size) -return -EREMOTEIO ; +if (!payload_size) { + ret = -EREMOTEIO ; + goto err; + }while (offset < payload_size) { tlv = (struct avs_tlv *)(payload + offset);
@@ -590,6 +597,9 @@ int avs_ipc_get_hw_config(struct avs_dev *adev, struct avs_hw_cfg *cfg) exit:/* No longer needed, free it as it's owned by the get_large_config() caller. */ kfree (payload); +err: +if (ret) +dev_err (adev->dev,"get hw cfg failed: %d\n" , ret);return ret; }
Signature · 1 line
2.25.1
Cezary Rojewski
maintainerPATCH 7/13 ASoC: Intel: avs: Add MODULE_FIRMWARE to inform about FW
In reply to Cezary RojewskiFrom: Amadeusz Sławiński <[email protected]>
MODULE_FIRMWARE macro adds hint to module information about which FW is expected to be present on file system.
--- sound/soc/intel/avs/core.c | 10 ++++++++++ 1 file changed, 10 insertions(+)
sound/soc/intel/avs/core.c +10 -0
@@ -902,3 +902,13 @@ MODULE_AUTHOR("Cezary Rojewski <[email protected]>");MODULE_AUTHOR ("Amadeusz Slawinski <[email protected]>" );MODULE_DESCRIPTION ("Intel cAVS sound driver" );MODULE_LICENSE ("GPL" ); +MODULE_FIRMWARE ("intel/skl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/apl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/cnl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/icl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/jsl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/lkf/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/tgl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/ehl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/adl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/adl_n/dsp_basefw.bin" );
Signature · 1 line
2.25.1
Pierre-Louis Bossart
reviewerRe: [PATCH 07/13] ASoC: Intel: avs: Add MODULE_FIRMWARE to inform about FW
sound/soc/intel/avs/core.c MODULE_AUTHOR("Cezary Rojewski <[email protected]>"); · line 910
Quoted 7 lines
sound/soc/intel/avs/core.c +6
-0+MODULE_FIRMWARE ("intel/skl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/apl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/cnl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/icl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/jsl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/lkf/dsp_basefw.bin" );
Lakefield? Is this really a supported platform? It was EOL'ed a year after launch, wasn't it?
sound/soc/intel/avs/core.c MODULE_AUTHOR("Cezary Rojewski <[email protected]>"); · line 914
Quoted 4 lines
sound/soc/intel/avs/core.c +4
-0+MODULE_FIRMWARE ("intel/tgl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/ehl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/adl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/adl_n/dsp_basefw.bin" );
If you start listing the variants of ADL, then shouldn't you also list the variants of TGL? Same for CNL, there are multiple variants, not to mention different signing keys.
Cezary Rojewski
maintainerRe: [PATCH 07/13] ASoC: Intel: avs: Add MODULE_FIRMWARE to inform about FW
Quoted 9 lines · Pierre-Louis Bossart
sound/soc/intel/avs/core.c MODULE_AUTHOR("Cezary Rojewski <[email protected]>"); · line 910
Quoted 7 lines
sound/soc/intel/avs/core.c +6
-0+MODULE_FIRMWARE ("intel/skl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/apl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/cnl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/icl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/jsl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/lkf/dsp_basefw.bin" );Lakefield? Is this really a supported platform? It was EOL'ed a year after launch, wasn't it?
We still keep all the platforms, regardless of their age, in the CI to cross check firmware-API compatibility even with newest generations such as PanterLake or FriscoLake.
Few years ago the team did quite a job to streamline LakeField (LKF) with other cAVS 2.5 platforms on the firmware side. The TLDR is:
- one firmware branch covers from LKF (cAVS 2.1) up to AlderLake/RaptorLake and all their spinoffs (cAVS 2.5)
- one tgl.c file on the avs-driver side covers the exact same range
so, any kind of redundancy is greatly limited. With that in mind, we utilize the cAVS CI against the ACE (MeteorLake onwards) equivalent to verify API compatibility between those two firmware branches to lower the software development cost.
Quoted 7 lines
sound/soc/intel/avs/core.c MODULE_AUTHOR("Cezary Rojewski <[email protected]>"); · line 914
Quoted 4 lines
sound/soc/intel/avs/core.c +4
-0+MODULE_FIRMWARE ("intel/tgl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/ehl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/adl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/adl_n/dsp_basefw.bin" );If you start listing the variants of ADL, then shouldn't you also list the variants of TGL? Same for CNL, there are multiple variants, not to mention different signing keys.
The only reason ADL and ADL-N are listed separately is binary incompatibility - MEU differs between the two. In essence, one can use ADL-binaries for all ADL/RPL based platforms _except_ for ADL-N based ones. The code is the same, the verification process is different. For all other major platforms, no MEU differences so one binary covers all variants.
In regard to the key, the approach is: it's ignored.
Whatever is in the directory under 'dsp_basefw.bin' will be attempted for booting the DSP. By default, what lands on the market is a production-type-signed binaries. Internally CI runs with prod -or- debug signed binaries but we do not intend to share the latter to the official linux-firmware repo.
Kind regards,
Czarek
Pierre-Louis Bossart
reviewerRe: [PATCH 07/13] ASoC: Intel: avs: Add MODULE_FIRMWARE to inform about FW
Quoted 9 lines
Quoted 8 lines
sound/soc/intel/avs/core.c MODULE_AUTHOR("Cezary Rojewski <[email protected]>"); · line 914
Quoted 5 lines
sound/soc/intel/avs/core.c +4
-0+MODULE_FIRMWARE ("intel/tgl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/ehl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/adl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/adl_n/dsp_basefw.bin" );If you start listing the variants of ADL, then shouldn't you also list the variants of TGL? Same for CNL, there are multiple variants, not to mention different signing keys.
Quoted 5 lines
The only reason ADL and ADL-N are listed separately is binary incompatibility - MEU differs between the two. In essence, one can use ADL-binaries for all ADL/RPL based platforms _except_ for ADL-N based ones. The code is the same, the verification process is different. For all other major platforms, no MEU differences so one binary covers all variants.
In regard to the key, the approach is: it's ignored.
Whatever is in the directory under 'dsp_basefw.bin' will be attempted for booting the DSP. By default, what lands on the market is a production-type-signed binaries. Internally CI runs with prod -or- debug signed binaries but we do not intend to share the latter to the official linux-firmware repo.
You didn't get my point. There are also binary incompatibilities between TGL and TGL-H (different number of cores) and in the CNL/CML space there were different production keys and MEU snaffus requiring different binaries to be released. This wasn't about the production/debug problem, I was only referring to 'production' releases.
Cezary Rojewski
maintainerRe: [PATCH 07/13] ASoC: Intel: avs: Add MODULE_FIRMWARE to inform about FW
Quoted 9 lines · Pierre-Louis Bossart
Quoted 9 lines
Quoted 8 lines
sound/soc/intel/avs/core.c MODULE_AUTHOR("Cezary Rojewski <[email protected]>"); · line 914
Quoted 5 lines
sound/soc/intel/avs/core.c +4
-0+MODULE_FIRMWARE ("intel/tgl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/ehl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/adl/dsp_basefw.bin" ); +MODULE_FIRMWARE ("intel/adl_n/dsp_basefw.bin" );If you start listing the variants of ADL, then shouldn't you also list the variants of TGL? Same for CNL, there are multiple variants, not to mention different signing keys.
Quoted 8 lines
Quoted 5 lines
The only reason ADL and ADL-N are listed separately is binary incompatibility - MEU differs between the two. In essence, one can use ADL-binaries for all ADL/RPL based platforms _except_ for ADL-N based ones. The code is the same, the verification process is different. For all other major platforms, no MEU differences so one binary covers all variants.
In regard to the key, the approach is: it's ignored.
Whatever is in the directory under 'dsp_basefw.bin' will be attempted for booting the DSP. By default, what lands on the market is a production-type-signed binaries. Internally CI runs with prod -or- debug signed binaries but we do not intend to share the latter to the official linux-firmware repo.
You didn't get my point. There are also binary incompatibilities between TGL and TGL-H (different number of cores) and in the CNL/CML space there were different production keys and MEU snaffus requiring different binaries to be released. This wasn't about the production/debug problem, I was only referring to 'production' releases.
I'm afraid there are differences between cAVS/ACE firmware and SOF. When it comes to the cAVS/ACE, TGL devices are all handled using the exact same binary and the firmware reads certain hardware capabilities in runtime. CML-R (CML platform with TGP PCH) falls into the same category.
Instead, the cAVS/ACE expects the software to send BUS_HARDWARE_ID IPC to inform it about details of the hardware revision e.g.: the stepping. Firmware may decide to tailor certain flows depending on that detailed information.
Kind regards,
Czarek
Cezary Rojewski
maintainerPATCH 8/13 ASoC: Intel: avs: Clearly state assumptions of hw_params()
In reply to Cezary RojewskiThere are no NULL-checks for fe/be_hw_params as there is an implicit assumption that framework opens valid DPCMs only. State that clearly.
--- sound/soc/intel/avs/pcm.c | 2 ++ 1 file changed, 2 insertions(+)
sound/soc/intel/avs/pcm.c +2 -0
@@ -161,6 +161,7 @@ static int avs_dai_be_hw_params(struct snd_pcm_substream *substream,struct snd_soc_dpcm *dpcm; be =snd_soc_substream_to_rtd (substream); +/* dpcm_fe_dai_open() guarantees the list is not empty at this point. */ for_each_dpcm_fe (be, substream->stream, dpcm) { fe = dpcm->fe; fe_hw_params = &fe->dpcm[substream->stream].hw_params;
@@ -576,6 +577,7 @@ static int avs_dai_fe_hw_params(struct snd_pcm_substream *substream,hdac_stream (host_stream)->format_val =0 ; fe =snd_soc_substream_to_rtd (substream); +/* dpcm_fe_dai_open() guarantees the list is not empty at this point. */ for_each_dpcm_be (fe, substream->stream, dpcm) { be = dpcm->be; be_hw_params = &be->dpcm[substream->stream].hw_params;
Signature · 1 line
2.25.1
Cezary Rojewski
maintainerPATCH 9/13 ASoC: Intel: avs: Improve logging of firmware loading
In reply to Cezary RojewskiCrucial debug information regarding the ROM/firmware status and last known error code is missing in the code loading functions.
--- sound/soc/intel/avs/loader.c | 26 ++++++++++++++++---------- 1 file changed, 16 insertions(+), 10 deletions(-)
sound/soc/intel/avs/loader.c +16 -10
@@ -167,7 +167,8 @@ int avs_cldma_load_basefw(struct avs_dev *adev, struct firmware *fw) (reg &AVS_ROM_INIT_DONE ) ==AVS_ROM_INIT_DONE , AVS_ROM_INIT_POLLING_US, SKL_ROM_INIT_TIMEOUT_US);if (ret <0 ) { -dev_err (adev->dev,"rom init timeout: %d\n" , ret); +dev_err (adev->dev,"rom init failed: %d, status: 0x%08x, lec: 0x%08x\n" , + ret, reg,snd_hdac_adsp_readl (adev,AVS_FW_REG_ERROR_CODE (adev)));avs_dsp_core_disable (adev,AVS_MAIN_CORE_MASK );return ret; }
@@ -180,7 +181,8 @@ int avs_cldma_load_basefw(struct avs_dev *adev, struct firmware *fw) AVS_FW_INIT_POLLING_US, AVS_FW_INIT_TIMEOUT_US);hda_cldma_stop (cl);if (ret <0 ) { -dev_err (adev->dev,"transfer fw failed: %d\n" , ret); +dev_err (adev->dev,"transfer fw failed: %d, status: 0x%08x, lec: 0x%08x\n" , + ret, reg,snd_hdac_adsp_readl (adev,AVS_FW_REG_ERROR_CODE (adev)));avs_dsp_core_disable (adev,AVS_MAIN_CORE_MASK );return ret; }
@@ -313,7 +315,8 @@ avs_hda_init_rom(struct avs_dev *adev, unsigned int dma_id, bool purge) (reg &0xF ) ==APL_ROM_FW_ENTERED , AVS_ROM_INIT_POLLING_US, APL_ROM_INIT_TIMEOUT_US);if (ret <0 ) { -dev_err (adev->dev,"rom init timeout: %d\n" , ret); +dev_err (adev->dev,"rom init failed: %d, status: 0x%08x, lec: 0x%08x\n" , + ret, reg,snd_hdac_adsp_readl (adev,AVS_FW_REG_ERROR_CODE (adev))); goto err; }
@@ -337,15 +340,15 @@ static int avs_imr_load_basefw(struct avs_dev *adev)/* DMA id ignored when flashing from IMR as no transfer occurs. */ ret =avs_hda_init_rom (adev,0 , false); -if (ret <0 ) { -dev_err (adev->dev,"rom init failed: %d\n" , ret); +if (ret <0 )return ret; - } ret =wait_for_completion_timeout (&adev->fw_ready,msecs_to_jiffies (AVS_FW_INIT_TIMEOUT_MS ));if (!ret) { -dev_err (adev->dev,"firmware ready timeout\n" ); +dev_err (adev->dev,"firmware ready timeout, status: 0x%08x, lec: 0x%08x\n" , +snd_hdac_adsp_readl (adev,AVS_FW_REG_STATUS (adev)), +snd_hdac_adsp_readl (adev,AVS_FW_REG_ERROR_CODE (adev)));avs_dsp_core_disable (adev,AVS_MAIN_CORE_MASK );return -ETIMEDOUT ; }
@@ -392,7 +395,7 @@ int avs_hda_load_basefw(struct avs_dev *adev, struct firmware *fw) ret =avs_hda_init_rom (adev, dma_id, true);if (!ret)break ; -dev_info (adev->dev,"#%d rom init fail: %d\n" , i +1 , ret); +dev_info (adev->dev,"#%d rom init failed: %d\n" , i +1 , ret); }if (ret <0 ) goto cleanup_resources;
@@ -404,7 +407,8 @@ int avs_hda_load_basefw(struct avs_dev *adev, struct firmware *fw) AVS_FW_INIT_POLLING_US, AVS_FW_INIT_TIMEOUT_US);snd_hdac_dsp_trigger (hstream, false);if (ret <0 ) { -dev_err (adev->dev,"transfer fw failed: %d\n" , ret); +dev_err (adev->dev,"transfer fw failed: %d, status: 0x%08x, lec: 0x%08x\n" , + ret, reg,snd_hdac_adsp_readl (adev,AVS_FW_REG_ERROR_CODE (adev)));avs_dsp_core_disable (adev,AVS_MAIN_CORE_MASK ); }
@@ -584,7 +588,9 @@ static int avs_dsp_load_basefw(struct avs_dev *adev) ret =wait_for_completion_timeout (&adev->fw_ready,msecs_to_jiffies (AVS_FW_INIT_TIMEOUT_MS ));if (!ret) { -dev_err (adev->dev,"firmware ready timeout\n" ); +dev_err (adev->dev,"firmware ready timeout, status: 0x%08x, lec: 0x%08x\n" , +snd_hdac_adsp_readl (adev,AVS_FW_REG_STATUS (adev)), +snd_hdac_adsp_readl (adev,AVS_FW_REG_ERROR_CODE (adev)));avs_dsp_core_disable (adev,AVS_MAIN_CORE_MASK ); ret = -ETIMEDOUT ; goto release_fw;
Signature · 1 line
2.25.1
Cezary Rojewski
maintainerPATCH 10/13 ASoC: Intel: avs: Update ASRC definition
In reply to Cezary RojewskiFrom: Amadeusz Sławiński <[email protected]>
To support ASRC for playback streams, update its descriptor.
--- sound/soc/intel/avs/messages.h | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)
sound/soc/intel/avs/messages.h +1 -2
@@ -859,8 +859,7 @@ static_assert(sizeof(struct avs_aec_cfg) == 92);struct avs_asrc_cfg {struct avs_modcfg_base base;u32 out_freq; -u32 rsvd0:1 ; -u32 mode:1 ; +u32 mode:2 ;u32 rsvd2:2 ;u32 disable_jitter_buffer:1 ;u32 rsvd3:27 ;
Signature · 1 line
2.25.1
Cezary Rojewski
maintainerPATCH 11/13 ASoC: Intel: avs: Adjust DSP status register names
In reply to Cezary RojewskiBoth status and error are "codes". Update the wording to make code cohesive.
--- sound/soc/intel/avs/loader.c | 12 ++++++------ sound/soc/intel/avs/registers.h | 2 +- 2 files changed, 7 insertions(+), 7 deletions(-)
sound/soc/intel/avs/loader.c +6 -6
@@ -168,7 +168,7 @@ int avs_cldma_load_basefw(struct avs_dev *adev, struct firmware *fw) AVS_ROM_INIT_POLLING_US, SKL_ROM_INIT_TIMEOUT_US);if (ret <0 ) {dev_err (adev->dev,"rom init failed: %d, status: 0x%08x, lec: 0x%08x\n" , - ret, reg,snd_hdac_adsp_readl (adev,AVS_FW_REG_ERROR_CODE (adev))); + ret, reg,snd_hdac_adsp_readl (adev,AVS_FW_REG_ERROR (adev)));avs_dsp_core_disable (adev,AVS_MAIN_CORE_MASK );return ret; }
@@ -182,7 +182,7 @@ int avs_cldma_load_basefw(struct avs_dev *adev, struct firmware *fw)hda_cldma_stop (cl);if (ret <0 ) {dev_err (adev->dev,"transfer fw failed: %d, status: 0x%08x, lec: 0x%08x\n" , - ret, reg,snd_hdac_adsp_readl (adev,AVS_FW_REG_ERROR_CODE (adev))); + ret, reg,snd_hdac_adsp_readl (adev,AVS_FW_REG_ERROR (adev)));avs_dsp_core_disable (adev,AVS_MAIN_CORE_MASK );return ret; }
@@ -316,7 +316,7 @@ avs_hda_init_rom(struct avs_dev *adev, unsigned int dma_id, bool purge) AVS_ROM_INIT_POLLING_US, APL_ROM_INIT_TIMEOUT_US);if (ret <0 ) {dev_err (adev->dev,"rom init failed: %d, status: 0x%08x, lec: 0x%08x\n" , - ret, reg,snd_hdac_adsp_readl (adev,AVS_FW_REG_ERROR_CODE (adev))); + ret, reg,snd_hdac_adsp_readl (adev,AVS_FW_REG_ERROR (adev))); goto err; }
@@ -348,7 +348,7 @@ static int avs_imr_load_basefw(struct avs_dev *adev)if (!ret) {dev_err (adev->dev,"firmware ready timeout, status: 0x%08x, lec: 0x%08x\n" ,snd_hdac_adsp_readl (adev,AVS_FW_REG_STATUS (adev)), -snd_hdac_adsp_readl (adev,AVS_FW_REG_ERROR_CODE (adev))); +snd_hdac_adsp_readl (adev,AVS_FW_REG_ERROR (adev)));avs_dsp_core_disable (adev,AVS_MAIN_CORE_MASK );return -ETIMEDOUT ; }
@@ -408,7 +408,7 @@ int avs_hda_load_basefw(struct avs_dev *adev, struct firmware *fw)snd_hdac_dsp_trigger (hstream, false);if (ret <0 ) {dev_err (adev->dev,"transfer fw failed: %d, status: 0x%08x, lec: 0x%08x\n" , - ret, reg,snd_hdac_adsp_readl (adev,AVS_FW_REG_ERROR_CODE (adev))); + ret, reg,snd_hdac_adsp_readl (adev,AVS_FW_REG_ERROR (adev)));avs_dsp_core_disable (adev,AVS_MAIN_CORE_MASK ); }
@@ -590,7 +590,7 @@ static int avs_dsp_load_basefw(struct avs_dev *adev)if (!ret) {dev_err (adev->dev,"firmware ready timeout, status: 0x%08x, lec: 0x%08x\n" ,snd_hdac_adsp_readl (adev,AVS_FW_REG_STATUS (adev)), -snd_hdac_adsp_readl (adev,AVS_FW_REG_ERROR_CODE (adev))); +snd_hdac_adsp_readl (adev,AVS_FW_REG_ERROR (adev)));avs_dsp_core_disable (adev,AVS_MAIN_CORE_MASK ); ret = -ETIMEDOUT ; goto release_fw;
sound/soc/intel/avs/registers.h +1 -1
@@ -74,7 +74,7 @@/* Constants used when accessing SRAM, space shared with firmware */ #define AVS_FW_REG_BASE (adev) ((adev)->spec->sram->base_offset)#define AVS_FW_REG_STATUS (adev) (AVS_FW_REG_BASE(adev) + 0x0) -#define AVS_FW_REG_ERROR_CODE (adev) (AVS_FW_REG_BASE(adev) + 0x4) +#define AVS_FW_REG_ERROR (adev) (AVS_FW_REG_BASE(adev) + 0x4)#define AVS_WINDOW_CHUNK_SIZE SZ_4K#define AVS_FW_REGS_SIZE AVS_WINDOW_CHUNK_SIZE
Signature · 1 line
2.25.1
Cezary Rojewski
maintainerPATCH 12/13 ASoC: Intel: avs: Adjust IPC traces
In reply to Cezary RojewskiThe firmware status and the firmware error registers are 4-bytes wide. Update trace macros and their usage to reflect that.
--- sound/soc/intel/avs/ipc.c | 25 ++++++++++++++---------- sound/soc/intel/avs/trace.h | 38 +++++++++++++++++++------------------ 2 files changed, 35 insertions(+), 28 deletions(-)
sound/soc/intel/avs/ipc.c +15 -10
@@ -184,10 +184,11 @@ static void avs_dsp_receive_rx(struct avs_dev *adev, u64 header) {struct avs_ipc *ipc = adev->ipc;union avs_reply_msg msg =AVS_MSG (header); -u64 reg; +u32 sts, lec; - reg =readq (avs_sram_addr (adev,AVS_FW_REGS_WINDOW )); -trace_avs_ipc_reply_msg (header, reg); + sts =snd_hdac_adsp_readl (adev,AVS_FW_REG_STATUS (adev)); + lec =snd_hdac_adsp_readl (adev,AVS_FW_REG_ERROR (adev)); +trace_avs_ipc_reply_msg (header, sts, lec); ipc->rx.header = header;/* Abort copying payload if request processing was unsuccessful. */
@@ -209,10 +210,11 @@ static void avs_dsp_process_notification(struct avs_dev *adev, u64 header)union avs_notify_msg msg =AVS_MSG (header);size_t data_size =0 ;void *data =NULL ; -u64 reg; +u32 sts, lec; - reg =readq (avs_sram_addr (adev,AVS_FW_REGS_WINDOW )); -trace_avs_ipc_notify_msg (header, reg); + sts =snd_hdac_adsp_readl (adev,AVS_FW_REG_STATUS (adev)); + lec =snd_hdac_adsp_readl (adev,AVS_FW_REG_ERROR (adev)); +trace_avs_ipc_notify_msg (header, sts, lec);/* Ignore spurious notifications until handshake is established. */ if (!adev->ipc->ready && msg.notify_msg_type !=AVS_NOTIFY_FW_READY ) {
@@ -367,13 +369,16 @@ static void avs_ipc_msg_init(struct avs_ipc *ipc, struct avs_ipc_msg *reply)static void avs_dsp_send_tx (struct avs_dev *adev,struct avs_ipc_msg *tx,bool read_fwregs) {const struct avs_spec *const spec = adev->spec; -u64 reg =ULONG_MAX ; +u32 sts =UINT_MAX ; +u32 lec =UINT_MAX ; tx->header |= spec->hipc->req_busy_mask; -if (read_fwregs) - reg =readq (avs_sram_addr (adev,AVS_FW_REGS_WINDOW )); +if (read_fwregs) { + sts =snd_hdac_adsp_readl (adev,AVS_FW_REG_STATUS (adev)); + lec =snd_hdac_adsp_readl (adev,AVS_FW_REG_ERROR (adev)); + } -trace_avs_request (tx, reg); +trace_avs_request (tx, sts, lec);if (tx->size)memcpy_toio (avs_downlink_addr (adev), tx->data, tx->size);
sound/soc/intel/avs/trace.h +20 -18
@@ -37,60 +37,62 @@ TRACE_EVENT(avs_dsp_core_op,void trace_avs_msg_payload (const void *data,size_t size); -#define trace_avs_request (msg, fwregs) \ +#define trace_avs_request (msg, sts, lec) \ ({ \ - trace_avs_ipc_request_msg((msg)->header, fwregs); \ + trace_avs_ipc_request_msg((msg)->header, sts, lec); \ trace_avs_msg_payload((msg)->data, (msg)->size); \ }) -#define trace_avs_reply (msg, fwregs) \ +#define trace_avs_reply (msg, sts, lec) \ ({ \ - trace_avs_ipc_reply_msg((msg)->header, fwregs); \ + trace_avs_ipc_reply_msg((msg)->header, sts, lec); \ trace_avs_msg_payload((msg)->data, (msg)->size); \ }) -#define trace_avs_notify (msg, fwregs) \ +#define trace_avs_notify (msg, sts, lec) \ ({ \ - trace_avs_ipc_notify_msg((msg)->header, fwregs); \ + trace_avs_ipc_notify_msg((msg)->header, sts, lec); \ trace_avs_msg_payload((msg)->data, (msg)->size); \ })#endif DECLARE_EVENT_CLASS (avs_ipc_msg_hdr, - TP_PROTO(u64 header, u64 fwregs), + TP_PROTO(u64 header, u32 sts, u32 lec), -TP_ARGS (header, fwregs), +TP_ARGS (header, sts, lec),TP_STRUCT__entry ( __field(u64, header) -__field (u64, fwregs) + __field(u32, sts) +__field (u32, lec) ),TP_fast_assign ( __entry->header = header; - __entry->fwregs = fwregs; + __entry->sts = sts; + __entry->lec = lec; ),TP_printk ("primary: 0x%08X, extension: 0x%08X,\n" -"fwstatus: 0x%08X, fwerror: 0x%08X" , +"status: 0x%08X, error: 0x%08X" ,lower_32_bits (__entry->header),upper_32_bits (__entry->header), -lower_32_bits (__entry->fwregs),upper_32_bits (__entry->fwregs)) + __entry->sts, __entry->lec) );DEFINE_EVENT (avs_ipc_msg_hdr, avs_ipc_request_msg, - TP_PROTO(u64 header, u64 fwregs), -TP_ARGS (header, fwregs) + TP_PROTO(u64 header, u32 sts, u32 lec), +TP_ARGS (header, sts, lec) );DEFINE_EVENT (avs_ipc_msg_hdr, avs_ipc_reply_msg, - TP_PROTO(u64 header, u64 fwregs), -TP_ARGS (header, fwregs) + TP_PROTO(u64 header, u32 sts, u32 lec), +TP_ARGS (header, sts, lec) );DEFINE_EVENT (avs_ipc_msg_hdr, avs_ipc_notify_msg, - TP_PROTO(u64 header, u64 fwregs), -TP_ARGS (header, fwregs) + TP_PROTO(u64 header, u32 sts, u32 lec), +TP_ARGS (header, sts, lec) ); TRACE_EVENT_CONDITION(avs_ipc_msg_payload,
Signature · 1 line
2.25.1
Cezary Rojewski
maintainerPATCH 13/13 ASoC: Intel: avs: Add missing includes
In reply to Cezary RojewskiThe debugfs file utilizes string helpers such as parse_int_array_user() yet does not include the header.
--- sound/soc/intel/avs/debugfs.c | 1 + 1 file changed, 1 insertion(+)
sound/soc/intel/avs/debugfs.c +1 -0
@@ -10,6 +10,7 @@#include <linux/kfifo.h> #include <linux/wait.h> #include <linux/sched/signal.h> +#include <linux/string_helpers.h> #include <sound/soc.h> #include "avs.h" #include "messages.h"
Signature · 1 line
2.25.1
Mark Brown
maintainerQuoted 10 lines · Cezary Rojewski
A set of loosely connected changes, fixing few outstanding issues as well as improving readability of the existing code.
The fixes lead the series, first five patches. The goal is to make sure proper
read()is used when accessing the registers,probe()andremove()sequences for HDAudio streaming are synced, minimal AudioDSP firmware version points to correct values and recent additions to the topology are parsed properly.[...]
Applied to
https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound.git for-next
Thanks!
[01/13] ASoC: Intel: avs: Do not readq() u32 registers
commit: bca0fa5f6b5e96c03daac1ed62b1e5c5057a2048
[02/13] ASoC: Intel: avs: Fix the minimum firmware version numbers
commit: dbda5c35b88794f6e5efe1b5b20044b0b3a340d4
[03/13] ASoC: Intel: avs: Fix theoretical infinite loop
commit: cf4d74256fe103ece7b2647550e6c063048e5682
[04/13] ASoC: Intel: avs: Fix init-config parsing
commit: e9ca3db9f01a7ce91ceab35cd5fa52f0c5aca174
[05/13] ASoC: Intel: avs: Update hda component teardown sequences
commit: e3146775f05d18c667a2e26082da3ac105f87d9f
[06/13] ASoC: Intel: avs: Print IPC error messages in lower layer
commit: 33228036ff655ebed1bc4bde9c9b6329b569b27b
[07/13] ASoC: Intel: avs: Add MODULE_FIRMWARE to inform about FW
commit: 94aa347d34e079859a5378272c9452c728e4183a
[08/13] ASoC: Intel: avs: Clearly state assumptions of hw_params()
commit: 0ca529926c5d9d0a3c0b1609fb7034ab870e4770
[09/13] ASoC: Intel: avs: Improve logging of firmware loading
commit: 480d9bb9cfb7b774b22cf82ff21903eb50d64cb9
[10/13] ASoC: Intel: avs: Update ASRC definition
commit: aea305d28551bc213aab3a41a0f59412247ae2b4
[11/13] ASoC: Intel: avs: Adjust DSP status register names
commit: 3eede0fc99c684df6f3f35866761036dabf89d05
[12/13] ASoC: Intel: avs: Adjust IPC traces
commit: ef724707788325a53ffa4cf58fceb94654e4793a
[13/13] ASoC: Intel: avs: Add missing includes
commit: 0b12850ddfb0032376ef1be10b5b46be00bba4d4All 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