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

[PATCH] ASoC: SOF: Intel: hda-loader: use small buffer for iccmax stream
by Peter Ujfalusi 10 Nov '22

10 Nov '22

From: Kai Vehmanen <kai.vehmanen(a)linux.intel.com> The data received via iccmax stream is not used anywhere, so no need to allocate a big DMA buffer for it. This is especially important as the allocation is done even in cases where reload of the firmware is skipped and execution happens directly from the firmware stored in IMR. Signed-off-by: Kai Vehmanen <kai.vehmanen(a)linux.intel.com> Reviewed-by: Péter Ujfalusi <peter.ujfalusi(a)linux.intel.com> Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart(a)linux.intel.com> Reviewed-by: Ranjani Sridharan <ranjani.sridharan(a)linux.intel.com> Signed-off-by: Peter Ujfalusi <peter.ujfalusi(a)linux.intel.com> --- sound/soc/sof/intel/hda-loader.c | 15 +++++---------- 1 file changed, 5 insertions(+), 10 deletions(-) diff --git a/sound/soc/sof/intel/hda-loader.c b/sound/soc/sof/intel/hda-loader.c index 3a4b0b6e2c5c..d680562edb35 100644 --- a/sound/soc/sof/intel/hda-loader.c +++ b/sound/soc/sof/intel/hda-loader.c @@ -322,7 +322,6 @@ int hda_dsp_cl_boot_firmware_iccmax(struct snd_sof_dev *sdev) { struct hdac_ext_stream *iccmax_stream; struct hdac_bus *bus = sof_to_bus(sdev); - struct firmware stripped_firmware; struct snd_dma_buffer dmab_bdl; int ret, ret1; u8 original_gb; @@ -330,15 +329,11 @@ int hda_dsp_cl_boot_firmware_iccmax(struct snd_sof_dev *sdev) /* save the original LTRP guardband value */ original_gb = snd_hdac_chip_readb(bus, VS_LTRP) & HDA_VS_INTEL_LTRP_GB_MASK; - if (sdev->basefw.fw->size <= sdev->basefw.payload_offset) { - dev_err(sdev->dev, "error: firmware size must be greater than firmware offset\n"); - return -EINVAL; - } - - stripped_firmware.size = sdev->basefw.fw->size - sdev->basefw.payload_offset; - - /* prepare capture stream for ICCMAX */ - iccmax_stream = hda_cl_stream_prepare(sdev, HDA_CL_STREAM_FORMAT, stripped_firmware.size, + /* + * Prepare capture stream for ICCMAX. We do not need to store + * the data, so use a buffer of PAGE_SIZE for receiving. + */ + iccmax_stream = hda_cl_stream_prepare(sdev, HDA_CL_STREAM_FORMAT, PAGE_SIZE, &dmab_bdl, SNDRV_PCM_STREAM_CAPTURE); if (IS_ERR(iccmax_stream)) { dev_err(sdev->dev, "error: dma prepare for ICCMAX stream failed\n"); -- 2.38.1

2 1

[PATCH 1/2] ASoC: mediatek: mt8183-da7219: Register to module device table
by Nícolas F. R. A. Prado 10 Nov '22

10 Nov '22

Register the compatibles for this module on the module device table so it can be automatically loaded when a matching device is found on the system. Signed-off-by: Nícolas F. R. A. Prado <nfraprado(a)collabora.com> --- sound/soc/mediatek/mt8183/mt8183-da7219-max98357.c | 1 + 1 file changed, 1 insertion(+) diff --git a/sound/soc/mediatek/mt8183/mt8183-da7219-max98357.c b/sound/soc/mediatek/mt8183/mt8183-da7219-max98357.c index 9f22d3939818..0e572fe28c58 100644 --- a/sound/soc/mediatek/mt8183/mt8183-da7219-max98357.c +++ b/sound/soc/mediatek/mt8183/mt8183-da7219-max98357.c @@ -842,6 +842,7 @@ static const struct of_device_id mt8183_da7219_max98357_dt_match[] = { }, {} }; +MODULE_DEVICE_TABLE(of, mt8183_da7219_max98357_dt_match); #endif static struct platform_driver mt8183_da7219_max98357_driver = { -- 2.38.1

3 5

[PATCH 0/3] ASoC: simpe-card/audio-graph fixups
by Kuninori Morimoto 10 Nov '22

10 Nov '22

Hi Mark These are simpe-card/audio-graph fixup patches. Kuninori Morimoto (3): ASoC: simple-card-utils: remove asoc_simple_convert_fixup() ASoC: audio-graph-card2-custom-sample.dtsi: add convert-rate for DPCM (MIXer) ASoC: audio-graph-card2: check also dpcm node for convert-xxx include/sound/simple_card_utils.h | 2 -- .../audio-graph-card2-custom-sample.dtsi | 10 ++++-- sound/soc/generic/audio-graph-card2.c | 3 +- sound/soc/generic/simple-card-utils.c | 35 +++++++------------ 4 files changed, 23 insertions(+), 27 deletions(-) -- 2.25.1 Thank you for your help !! Best regards --- Kuninori Morimoto

2 4

[PATCH 0/7] ASoC: codecs: da7219: Do not export da7219_aad_jack_det()
by Cezary Rojewski 10 Nov '22

10 Nov '22

Goal of the series is to drop any usage of the internal function. First, update all users to utilize component->set_jack() and then remove the relevant EXPORT_SYMBOL_GPL. Amadeusz Sławiński (2): ASoC: codecs: da7219: Introduce set_jack() callback ASoC: codecs: da7219: Do not export internal symbols Cezary Rojewski (5): ASoC: Intel: Drop da7219_aad_jack_det() usage ASoC: mediatek: Drop da7219_aad_jack_det() usage ASoC: rockchip: Drop da7219_aad_jack_det() usage ASoC: amd: Drop da7219_aad_jack_det() usage ASoC: Intel: avs: Drop da7219_aad_jack_det() usage sound/soc/amd/acp-da7219-max98357a.c | 3 +-- sound/soc/codecs/da7219-aad.c | 4 ---- sound/soc/codecs/da7219.c | 9 +++++++++ sound/soc/intel/avs/boards/da7219.c | 3 +-- sound/soc/intel/boards/bxt_da7219_max98357a.c | 3 +-- sound/soc/intel/boards/kbl_da7219_max98357a.c | 3 +-- sound/soc/intel/boards/kbl_da7219_max98927.c | 3 +-- sound/soc/intel/boards/sof_da7219_max98373.c | 4 ++-- sound/soc/mediatek/mt8183/mt8183-da7219-max98357.c | 3 +-- .../soc/mediatek/mt8186/mt8186-mt6366-da7219-max98357.c | 4 ++-- sound/soc/rockchip/rk3399_gru_sound.c | 3 +-- 11 files changed, 20 insertions(+), 22 deletions(-) -- 2.25.1

2 8

[PATCH linux-next v2] ASoC: SOF: remove duplicated included sof-audio.h
by yang.yang29＠zte.com.cn 10 Nov '22

10 Nov '22

From: Xu Panda <xu.panda(a)zte.com.cn> The sof-audio.h is included more than once. Reported-by: Zeal Robot <zealci(a)zte.com.cn> Signed-off-by: Xu Panda <xu.panda(a)zte.com.cn> Signed-off-by: Yang Yang <yang.yang29(a)zte.com> Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart(a)linux.intel.com> --- change for v2 - add maintainers and the alsa-devel mailing list in CC. --- sound/soc/sof/amd/acp-common.c | 1 - 1 file changed, 1 deletion(-) diff --git a/sound/soc/sof/amd/acp-common.c b/sound/soc/sof/amd/acp-common.c index 27b95187356e..348e70dfe2a5 100644 --- a/sound/soc/sof/amd/acp-common.c +++ b/sound/soc/sof/amd/acp-common.c @@ -13,7 +13,6 @@ #include "../sof-priv.h" #include "../sof-audio.h" #include "../ops.h" -#include "../sof-audio.h" #include "acp.h" #include "acp-dsp-offset.h" -- 2.15.2

3 2

[PATCH 0/2] ASoC: Intel: avs: DSP recovery and resume fixes
by Cezary Rojewski 10 Nov '22

10 Nov '22

Two fixes that are result of the recent discussions [1][2]. First adds missing locking around snd_pcm_stop() while the second fix sets substream state to DISCONNECTED if any suspend/resume related operation fails so that userspace has means to be aware that something went wrong during said operation. [1]: https://lore.kernel.org/alsa-devel/73e6425f-8e51-e26f-ad83-ccc5df517a43@int… [2]: https://lore.kernel.org/alsa-devel/20221104131244.3920179-1-cezary.rojewski… Cezary Rojewski (2): ASoC: Intel: avs: Lock substream before snd_pcm_stop() ASoC: Intel: avs: Disconnect substream if suspend or resume fails sound/soc/intel/avs/ipc.c | 2 ++ sound/soc/intel/avs/pcm.c | 10 ++++++++-- 2 files changed, 10 insertions(+), 2 deletions(-) -- 2.25.1

2 8

[PATCH] ALSA: hda: fix potential memleak in 'add_widget_node'
by Ye Bin 10 Nov '22

10 Nov '22

From: Ye Bin <yebin10(a)huawei.com> As 'kobject_add' may allocated memory for 'kobject->name' when return error. And in this function, if call 'kobject_add' failed didn't free kobject. So call 'kobject_put' to recycling resources. Signed-off-by: Ye Bin <yebin10(a)huawei.com> --- sound/hda/hdac_sysfs.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/sound/hda/hdac_sysfs.c b/sound/hda/hdac_sysfs.c index e47de49a32e3..62a9615dcf52 100644 --- a/sound/hda/hdac_sysfs.c +++ b/sound/hda/hdac_sysfs.c @@ -346,8 +346,10 @@ static int add_widget_node(struct kobject *parent, hda_nid_t nid, return -ENOMEM; kobject_init(kobj, &widget_ktype); err = kobject_add(kobj, parent, "%02x", nid); - if (err < 0) + if (err < 0) { + kobject_put(kobj); return err; + } err = sysfs_create_group(kobj, group); if (err < 0) { kobject_put(kobj); -- 2.31.1

2 1

[PATCH] ASoC: codecs: wsa883x: Use proper shutdown GPIO values
by Krzysztof Kozlowski 10 Nov '22

10 Nov '22

The shutdown GPIO is active_low (sd_n), but this depends on actual board layout. Linux drivers should only care about logical state, where high (1) means shutdown and low (0) means do not shutdown. Invert the GPIO to match logical value. Fixes: 43b8c7dc85a1 ("ASoC: codecs: add wsa883x amplifier support") Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski(a)linaro.org> --- sound/soc/codecs/wsa883x.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/sound/soc/codecs/wsa883x.c b/sound/soc/codecs/wsa883x.c index c7b10bbfba7e..77a7dd3cf495 100644 --- a/sound/soc/codecs/wsa883x.c +++ b/sound/soc/codecs/wsa883x.c @@ -1392,7 +1392,7 @@ static int wsa883x_probe(struct sdw_slave *pdev, } wsa883x->sd_n = devm_gpiod_get_optional(&pdev->dev, "powerdown", - GPIOD_FLAGS_BIT_NONEXCLUSIVE); + GPIOD_FLAGS_BIT_NONEXCLUSIVE | GPIOD_OUT_HIGH); if (IS_ERR(wsa883x->sd_n)) { dev_err(&pdev->dev, "Shutdown Control GPIO not found\n"); ret = PTR_ERR(wsa883x->sd_n); @@ -1411,7 +1411,7 @@ static int wsa883x_probe(struct sdw_slave *pdev, pdev->prop.simple_clk_stop_capable = true; pdev->prop.sink_dpn_prop = wsa_sink_dpn_prop; pdev->prop.scp_int1_mask = SDW_SCP_INT1_BUS_CLASH | SDW_SCP_INT1_PARITY; - gpiod_direction_output(wsa883x->sd_n, 1); + gpiod_direction_output(wsa883x->sd_n, 0); wsa883x->regmap = devm_regmap_init_sdw(pdev, &wsa883x_regmap_config); if (IS_ERR(wsa883x->regmap)) { -- 2.34.1

1 1

[PATCH] ALSA: memalloc: Don't fall back for SG-buffer with IOMMU
by Takashi Iwai 10 Nov '22

10 Nov '22

When the non-contiguous page allocation for SG buffer allocation fails, the memalloc helper tries to fall back to the old page allocation methods. This would, however, result in the bogus page addresses when IOMMU is enabled. Usually in such a case, the fallback allocation should fail as well, but occasionally it succeeds and hitting a bad access. The fallback was thought for non-IOMMU case, and as the error from dma_alloc_noncontiguous() with IOMMU essentially implies a fatal memory allocation error, we should return the error straightforwardly without fallback. This avoids the corner case like the above. The patch also renames the local variable "dma_ops" with snd_ prefix for avoiding the name conflict. Fixes: a8d302a0b770 ("ALSA: memalloc: Revive x86-specific WC page allocations again") Reported-by: Kai Vehmanen <kai.vehmanen(a)linux.intel.com> Reviewed-by: Kai Vehmanen <kai.vehmanen(a)linux.intel.com> Link: https://lore.kernel.org/r/alpine.DEB.2.22.394.2211041541090.3532114@elitele… Signed-off-by: Takashi Iwai <tiwai(a)suse.de> --- sound/core/memalloc.c | 20 +++++++++++--------- 1 file changed, 11 insertions(+), 9 deletions(-) diff --git a/sound/core/memalloc.c b/sound/core/memalloc.c index 03cffe771366..6a81aaab25ab 100644 --- a/sound/core/memalloc.c +++ b/sound/core/memalloc.c @@ -9,6 +9,7 @@ #include <linux/slab.h> #include <linux/mm.h> #include <linux/dma-mapping.h> +#include <linux/dma-map-ops.h> #include <linux/genalloc.h> #include <linux/highmem.h> #include <linux/vmalloc.h> @@ -541,19 +542,20 @@ static void *snd_dma_noncontig_alloc(struct snd_dma_buffer *dmab, size_t size) struct sg_table *sgt; void *p; - sgt = dma_alloc_noncontiguous(dmab->dev.dev, size, dmab->dev.dir, - DEFAULT_GFP, 0); - if (!sgt) { #ifdef CONFIG_SND_DMA_SGBUF + if (!get_dma_ops(dmab->dev.dev)) { if (dmab->dev.type == SNDRV_DMA_TYPE_DEV_WC_SG) dmab->dev.type = SNDRV_DMA_TYPE_DEV_WC_SG_FALLBACK; else dmab->dev.type = SNDRV_DMA_TYPE_DEV_SG_FALLBACK; return snd_dma_sg_fallback_alloc(dmab, size); -#else - return NULL; -#endif } +#endif + + sgt = dma_alloc_noncontiguous(dmab->dev.dev, size, dmab->dev.dir, + DEFAULT_GFP, 0); + if (!sgt) + return NULL; dmab->dev.need_sync = dma_need_sync(dmab->dev.dev, sg_dma_address(sgt->sgl)); @@ -857,7 +859,7 @@ static const struct snd_malloc_ops snd_dma_noncoherent_ops = { /* * Entry points */ -static const struct snd_malloc_ops *dma_ops[] = { +static const struct snd_malloc_ops *snd_dma_ops[] = { [SNDRV_DMA_TYPE_CONTINUOUS] = &snd_dma_continuous_ops, [SNDRV_DMA_TYPE_VMALLOC] = &snd_dma_vmalloc_ops, #ifdef CONFIG_HAS_DMA @@ -883,7 +885,7 @@ static const struct snd_malloc_ops *snd_dma_get_ops(struct snd_dma_buffer *dmab) if (WARN_ON_ONCE(!dmab)) return NULL; if (WARN_ON_ONCE(dmab->dev.type <= SNDRV_DMA_TYPE_UNKNOWN || - dmab->dev.type >= ARRAY_SIZE(dma_ops))) + dmab->dev.type >= ARRAY_SIZE(snd_dma_ops))) return NULL; - return dma_ops[dmab->dev.type]; + return snd_dma_ops[dmab->dev.type]; } -- 2.35.3

1 0

Re: regression with SG DMA buf allocations and IOMMU in low-mem
by Kai Vehmanen 10 Nov '22

10 Nov '22

Hi, [ and sorry, I had the bounces address in cc originally, routing reply to list as well ] On Fri, 4 Nov 2022, Takashi Iwai wrote: > On Fri, 04 Nov 2022 15:56:50 +0100, Kai Vehmanen wrote: > > I've been debugging a SOF DSP load fail on VT-D/IOMMU systems [1] and > > that brought me to these two commits from you: > > .. but in rare low-memory cases, the dma_alloc_noncontiguous() > > call will fail in core/memalloc.c and we go to the fallback path. > > > > So we go to snd_dma_sg_fallback_alloc(), but here it seems something is > > missing. The pages are not allocated with DMA mapping API anymore, so > > IOMMU won't know about the memory and in our case the DSP load will fail > > to a IOMMU fault. [...] > > Hrm, that's a tough issue. Basically if dma_alloc_noncontiguous() > fails, it's really out of memory -- at least under the situation where > IOMMU is enabled. The fallback path was introduced for the case where > there is no IOMMU and noncontiguous allocation fails from the > beginning. ack. This matches with the bug reports. These happen rarely and on systems with a lot of memory pressure. We have recently submitted a few features&fixes that will further reduce these allocs, so it probably is better outcome to have a plain out of memory error. > And, what makes things more complicated is that snd_dma_dev_alloc() > cannot be used for SG pages when IOMMU is involved. We can't get the > proper pages for mapping SG from the DMA address in that case; some > systems may work with virt_to_page() but it's not guranteed to work. Aa, ok, so we need to use dma_alloc_noncontiguous() to do this properly and if it returns ENOMEM, that's what it is then, right. > So, I believe there are two issues: > 1. The firmware allocation fails with dma_alloc_noncontiguous() with > IOMMU enabled > > 2. The helper goes to the fallback path and occasionally it worked, > although the pages can't be used with IOMMU. > > Basically when 1 happens, we should just return an error without going > to fallback path. But it won't "fix" your case? I think an explicit error would be best. The problem now is that the driver will think the allocation (and mapping to device) is fine and proceeds to program the hardware to use the address. This will then create an IOMMU fault down the line that is not so straighforward to recover from (worst case is that a full device level reset needs to be done). And given driver doesn't know it got a faulty mapping, it's hard to make the decision why the fault happened. Br, Kai

2 3