April 2022 - Alsa-devel - mailman.alsa-project.org

[PATCH] ASoC: SOF: Kconfig: Make SND_SOC_SOF_HDA_PROBES tristate
by Richard Fitzgerald 21 Apr '22

21 Apr '22

SND_SOC_SOF_HDA_PROBES must be tristate because the code it builds depends on code that is tristate. If SND_SOC_SOF_HDA_PROBES is bool it leads to the following build inconsistency: SND_SOC_SOF_HDA_COMMON=m which selects SND_SOC_SOF_HDA_PROBES but since this is a bool SND_SOC_SOF_HDA_PROBES=y SND_SOC_SOF_HDA_PROBES=y selects SND_SOC_SOF_DEBUG_PROBES=y so sof-client-probes.c is built into the kernel. sof-client-probes.c calls functions in sof-client.c, but SND_SOC_SOF=m sof-client.c is built into a loadable module. Signed-off-by: Richard Fitzgerald <rf(a)opensource.cirrus.com> --- sound/soc/sof/intel/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/soc/sof/intel/Kconfig b/sound/soc/sof/intel/Kconfig index 172419392b33..0def2aa5581d 100644 --- a/sound/soc/sof/intel/Kconfig +++ b/sound/soc/sof/intel/Kconfig @@ -260,7 +260,7 @@ config SND_SOC_SOF_HDA 'select' statements at a higher level. config SND_SOC_SOF_HDA_PROBES - bool + tristate select SND_SOC_SOF_DEBUG_PROBES help The option enables the data probing for Intel(R) Skylake and newer -- 2.30.2

2 1

[PATCH] ASoC: dmic: Add support for DSD data format
by Shengjiu Wang 21 Apr '22

21 Apr '22

Add DSD format support in this generic dmic driver: DSD_U8, DSD_U16_LE, DSD_U32_LE, Signed-off-by: Shengjiu Wang <shengjiu.wang(a)nxp.com> --- sound/soc/codecs/dmic.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/sound/soc/codecs/dmic.c b/sound/soc/codecs/dmic.c index 5d079d90fd3b..d1a30ca4571a 100644 --- a/sound/soc/codecs/dmic.c +++ b/sound/soc/codecs/dmic.c @@ -82,7 +82,10 @@ static struct snd_soc_dai_driver dmic_dai = { .rates = SNDRV_PCM_RATE_CONTINUOUS, .formats = SNDRV_PCM_FMTBIT_S32_LE | SNDRV_PCM_FMTBIT_S24_LE - | SNDRV_PCM_FMTBIT_S16_LE, + | SNDRV_PCM_FMTBIT_S16_LE + | SNDRV_PCM_FMTBIT_DSD_U8 + | SNDRV_PCM_FMTBIT_DSD_U16_LE + | SNDRV_PCM_FMTBIT_DSD_U32_LE, }, .ops = &dmic_dai_ops, }; -- 2.17.1

2 1

[PATCH v6 00/21] ASoC: fsl_micfil: Driver updates
by Sascha Hauer 21 Apr '22

21 Apr '22

I added one more ack from Shengjiu Wang, but other than that it's just a resend with +Cc: Mark Brown Sascha Changes since v5: - Add one more ack from Shengjiu Wang Changes since v4: - collect more acks - whitespace cleanup in 16/21 Changes since v3: - Add commit log to "ASoC: fsl_micfil: drop unused variables" - Fix include name in "ASoC: fsl_micfil: add multi fifo support" - Drop unnecessary temporary adding of struct fsl_micfil::osr - Leave default quality setting at 'medium' - Drop debugging message printed at error level - collect acks from Shengjiu Wang and Vinod Koul Changes since v2: - Add forgotten commit log to dmaengine patches - Add patch to move include/linux/platform_data/dma-imx.h to include/linux/dma/imx-dma.h - Use prefix dmaengine: for dma patches Changes since v1: - Drop unused variable sw_done_sel - Evaluate sdmac->direction directly instead of storing value in n_fifos - add missing include linux/bitfield.h Sascha Hauer (21): ASoC: fsl_micfil: Drop unnecessary register read ASoC: fsl_micfil: Drop unused register read ASoC: fsl_micfil: drop fsl_micfil_set_mclk_rate() ASoC: fsl_micfil: do not define SHIFT/MASK for single bits ASoC: fsl_micfil: use GENMASK to define register bit fields ASoC: fsl_micfil: use clear/set bits ASoC: fsl_micfil: drop error messages from failed register accesses ASoC: fsl_micfil: drop unused variables dmaengine: imx: Move header to include/dma/ dmaengine: imx-sdma: error out on unsupported transfer types dmaengine: imx-sdma: Add multi fifo support ASoC: fsl_micfil: add multi fifo support ASoC: fsl_micfil: use define for OSR default value ASoC: fsl_micfil: Drop get_pdm_clk() ASoC: fsl_micfil: simplify clock setting ASoC: fsl_micfil: rework quality setting ASoC: fsl_micfil: drop unused include ASoC: fsl_micfil: drop only once used defines ASoC: fsl_micfil: drop support for undocumented property ASoC: fsl_micfil: fold fsl_set_clock_params() into its only user ASoC: fsl_micfil: Remove debug message drivers/dma/imx-dma.c | 2 +- drivers/dma/imx-sdma.c | 76 +++- drivers/mmc/host/mxcmmc.c | 2 +- drivers/spi/spi-fsl-lpspi.c | 2 +- drivers/spi/spi-imx.c | 2 +- drivers/tty/serial/imx.c | 2 +- drivers/video/fbdev/mx3fb.c | 2 +- .../dma-imx.h => dma/imx-dma.h} | 26 +- sound/soc/fsl/fsl_asrc.c | 2 +- sound/soc/fsl/fsl_asrc_dma.c | 2 +- sound/soc/fsl/fsl_easrc.h | 2 +- sound/soc/fsl/fsl_micfil.c | 369 +++++++----------- sound/soc/fsl/fsl_micfil.h | 269 +++---------- sound/soc/fsl/imx-pcm.h | 2 +- sound/soc/fsl/imx-ssi.h | 2 +- 15 files changed, 297 insertions(+), 465 deletions(-) rename include/linux/{platform_data/dma-imx.h => dma/imx-dma.h} (67%) -- 2.30.2

2 22

[PATCH v2 00/15] ASoC: SOF: add INTEL_IPC4 plumbing
by Pierre-Louis Bossart 21 Apr '22

21 Apr '22

The INTEL_IPC4 protocol and firmware architecture will rely on different sets of firmware binary and topology files. Some platforms will only support INTEL_IPC4, some will support both INTEL_IPC4 and SOF_IPC for development, and some will stay with the existing SOF_IPC. This patchset adds new IPC definitions, and search paths for firmware and topology files, along with means to override the default IPC type and search paths for development. The firmware binary names are aligned with those used by the Intel AVS driver to avoid duplicate firmware installs, but the topology will have to differ due to driver architecture differences. Changes since v1: Removed SKL/KBL code loader and initial hardware support. This will be re-submitted as a follow-up patchset. The rest of the code is 100% identical to v1. Pierre-Louis Bossart (11): ASoC: SOF: add definitions for different IPC types ASoC: SOF: add path indirection to each IPC type ASoC: SOF: add default IPC capability and file paths ASoC: SOF: add IPC-dependent file names ASoC: SOF: sof-pci-dev: add parameter to override firmware filename ASoC: SOF: sof-pci-dev: add parameter to override topology filename ASoC: SOF: sof-audio: preserve topology filename override in nocodec mode ASoC: SOF: remove const qualifier for 'struct snd_sof_dsp_ops' ASoC: SOF: Intel: hda: use common ops across platforms ASoC: SOF: Intel: hda: use BIT() macros for consistency ASoC: SOF: Intel: add IP identifier Ranjani Sridharan (4): ASoC: SOF: Intel: move HDA_CL_STREAM_FORMAT definition to header ASoC: SOF: Intel: hda: define check_ipc_irq op ASoC: SOF: Intel: hda: Define rom_status_reg in sof_intel_dsp_desc ASoC: SOF: Intel: hda: expose some codeloader functions include/sound/sof.h | 23 ++++- sound/soc/sof/amd/acp.h | 2 +- sound/soc/sof/amd/pci-rn.c | 14 ++- sound/soc/sof/amd/renoir.c | 2 +- sound/soc/sof/core.c | 6 ++ sound/soc/sof/debug.c | 2 +- sound/soc/sof/imx/imx8.c | 32 +++++-- sound/soc/sof/imx/imx8m.c | 16 +++- sound/soc/sof/intel/Makefile | 2 +- sound/soc/sof/intel/apl.c | 107 +++++----------------- sound/soc/sof/intel/bdw.c | 17 +++- sound/soc/sof/intel/byt.c | 48 +++++++--- sound/soc/sof/intel/cnl.c | 112 ++++++----------------- sound/soc/sof/intel/hda-common-ops.c | 107 ++++++++++++++++++++++ sound/soc/sof/intel/hda-loader.c | 43 ++++----- sound/soc/sof/intel/hda.c | 23 ++++- sound/soc/sof/intel/hda.h | 30 +++++-- sound/soc/sof/intel/icl.c | 111 +++++------------------ sound/soc/sof/intel/pci-apl.c | 36 ++++++-- sound/soc/sof/intel/pci-cnl.c | 54 +++++++++-- sound/soc/sof/intel/pci-icl.c | 36 ++++++-- sound/soc/sof/intel/pci-tgl.c | 91 +++++++++++++++---- sound/soc/sof/intel/pci-tng.c | 17 +++- sound/soc/sof/intel/shim.h | 16 +++- sound/soc/sof/intel/tgl.c | 120 +++++++------------------ sound/soc/sof/mediatek/mt8195/mt8195.c | 16 +++- sound/soc/sof/ops.h | 8 ++ sound/soc/sof/pcm.c | 2 +- sound/soc/sof/sof-acpi-dev.c | 6 +- sound/soc/sof/sof-audio.c | 3 +- sound/soc/sof/sof-of-dev.c | 6 +- sound/soc/sof/sof-pci-dev.c | 47 ++++++++-- 32 files changed, 671 insertions(+), 484 deletions(-) create mode 100644 sound/soc/sof/intel/hda-common-ops.c -- 2.30.2

2 16

[PATCH v3 0/3] soundwire: qcom: add pm runtime support
by Srinivas Kandagatla 20 Apr '22

20 Apr '22

This patchset adds pm runtime support to Qualcomm SounWire Controller using SoundWire Clock Stop and Wake up using Headset events on supported instances and instances like WSA which do not support clock stop a soft reset of controller along with full rest of slaves is done to resume from a low power state. Tested it on SM8250 MTP and Dragon Board DB845c Changes since v2: - update log as suggested by Pierre - removing handling clk stop for cases where the controller is soft reset. - add more error checks when calling sdw_bus_prep_clk_stop - update dt-bindings with wakeup-source and interrupt-names properties. --srini Srinivas Kandagatla (3): soundwire: qcom: add runtime pm support dt-bindings: soundwire: qcom: document optional wake irq soundwire: qcom: add in-band wake up interrupt support .../bindings/soundwire/qcom,sdw.txt | 14 +- drivers/soundwire/qcom.c | 206 +++++++++++++++++- 2 files changed, 218 insertions(+), 2 deletions(-) -- 2.21.0

5 15

[PATCH 00/18] xen: simplify frontend side ring setup
by Juergen Gross 20 Apr '22

20 Apr '22

Many Xen PV frontends share similar code for setting up a ring page (allocating and granting access for the backend) and for tearing it down. Create new service functions doing all needed steps in one go. This requires all frontends to use a common value for an invalid grant reference in order to make the functions idempotent. Juergen Gross (18): xen/blkfront: switch blkfront to use INVALID_GRANT_REF xen/netfront: switch netfront to use INVALID_GRANT_REF xen/scsifront: remove unused GRANT_INVALID_REF definition xen/usb: switch xen-hcd to use INVALID_GRANT_REF xen/drm: switch xen_drm_front to use INVALID_GRANT_REF xen/sound: switch xen_snd_front to use INVALID_GRANT_REF xen/dmabuf: switch gntdev-dmabuf to use INVALID_GRANT_REF xen/shbuf: switch xen-front-pgdir-shbuf to use INVALID_GRANT_REF xen/xenbus: add xenbus_setup_ring() service function xen/blkfront: use xenbus_setup_ring() and xenbus_teardown_ring() xen/netfront: use xenbus_setup_ring() and xenbus_teardown_ring() xen/tpmfront: use xenbus_setup_ring() and xenbus_teardown_ring() xen/drmfront: use xenbus_setup_ring() and xenbus_teardown_ring() xen/pcifront: use xenbus_setup_ring() and xenbus_teardown_ring() xen/scsifront: use xenbus_setup_ring() and xenbus_teardown_ring() xen/usbfront: use xenbus_setup_ring() and xenbus_teardown_ring() xen/sndfront: use xenbus_setup_ring() and xenbus_teardown_ring() xen/xenbus: eliminate xenbus_grant_ring() drivers/block/xen-blkfront.c | 54 ++++---------- drivers/char/tpm/xen-tpmfront.c | 18 +---- drivers/gpu/drm/xen/xen_drm_front.h | 9 --- drivers/gpu/drm/xen/xen_drm_front_evtchnl.c | 40 +++------- drivers/net/xen-netfront.c | 77 ++++++-------------- drivers/pci/xen-pcifront.c | 19 +---- drivers/scsi/xen-scsifront.c | 30 ++------ drivers/usb/host/xen-hcd.c | 59 ++++----------- drivers/xen/gntdev-dmabuf.c | 13 +--- drivers/xen/xen-front-pgdir-shbuf.c | 17 +---- drivers/xen/xenbus/xenbus_client.c | 81 ++++++++++++++++----- include/xen/xenbus.h | 4 +- sound/xen/xen_snd_front_evtchnl.c | 41 +++-------- sound/xen/xen_snd_front_evtchnl.h | 9 --- 14 files changed, 156 insertions(+), 315 deletions(-) -- 2.34.1

1 2

[PATCH] ASoC: fsl: Fix error handling in pcm030_fabric_probe
by Miaoqian Lin 20 Apr '22

20 Apr '22

This will call twice platform_device_put() if both platform_device_add() and snd_soc_register_card() fails. Fix it by using goto label to avoid duplicating the error code logic. Fixes: fb25621da570 ("ASoC: fsl: Add missing error handling in pcm030_fabric_probe") Signed-off-by: Miaoqian Lin <linmq006(a)gmail.com> --- sound/soc/fsl/pcm030-audio-fabric.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) diff --git a/sound/soc/fsl/pcm030-audio-fabric.c b/sound/soc/fsl/pcm030-audio-fabric.c index 83b4a22bf15a..d397bb97f37b 100644 --- a/sound/soc/fsl/pcm030-audio-fabric.c +++ b/sound/soc/fsl/pcm030-audio-fabric.c @@ -95,19 +95,23 @@ static int pcm030_fabric_probe(struct platform_device *op) ret = platform_device_add(pdata->codec_device); if (ret) { dev_err(&op->dev, "platform_device_add() failed: %d\n", ret); - platform_device_put(pdata->codec_device); + goto err_add; } ret = snd_soc_register_card(card); if (ret) { dev_err(&op->dev, "snd_soc_register_card() failed: %d\n", ret); - platform_device_del(pdata->codec_device); - platform_device_put(pdata->codec_device); + goto err_register; } platform_set_drvdata(op, pdata); return ret; +err_register: + platform_device_del(pdata->codec_device); +err_add: + platform_device_put(pdata->codec_device); + return ret; } static int pcm030_fabric_remove(struct platform_device *op) -- 2.17.1

2 6

Proposal: MIDI instead of input device for Caiaq devices
by Simon Larcher 20 Apr '22

20 Apr '22

Hi, I'd like to suggest changing the snd-usb-caiaq driver module to send MIDI messages instead of emulating an input device. This proposal more precisely refers to the Native Instruments Rig Kontrol 3 which I personally own. It's a 2in/2out USB sound card in a pedalboard housing with 8 footswitches and an expression pedal. To me it makes more sense to use MIDI as the device is built for that purpose. The configuration software provided by the manufacturer obviously doesn't work on Linux; but what it does is mapping the physical switches to MIDI messages. On linux, as is the device emulates an input device that sends 1-9 keystrokes for the footswitches and reports a 3 axis joystick for the onboard pedal + the 2 external pedal inputs. The onboard leds and 4 digits 7-segments display are available as controls (as listed by eg. amixer -c0 controls), this could also be changed to using MIDI as well. I am hoping for this to be approved and am open to suggestions regarding the implementation details. More specifically if it makes more sense to use Note On/Off messages or Controller Change messages, as well as the naming for the rawmidi inputs. The device will have two pairs of MIDI ports, one for the external I/O MIDI connectors and one for the "virtual" input reporting the button presses. I have already started working on it and I have a very rough POC of this proposal, adding an extra MIDI device and sending a CC message when the switches are pressed. It is available here: https://framagit.org/Simon-L/snd-usb-caiaq This has been discussed on #lad @ Libera.chat, I am registered as FergusL there. Thanks! SL

1 0

[PATCH] ALSA: usb-audio: Clear MIDI port active flag after draining
by Takashi Iwai 20 Apr '22

20 Apr '22

When a rawmidi output stream is closed, it calls the drain at first, then does trigger-off only when the drain returns -ERESTARTSYS as a fallback. It implies that each driver should turn off the stream properly after the drain. Meanwhile, USB-audio MIDI interface didn't change the port->active flag after the drain. This may leave the output work picking up the port that is closed right now, which eventually leads to a use-after-free for the already released rawmidi object. This patch fixes the bug by properly clearing the port->active flag after the output drain. Reported-by: syzbot+70e777a39907d6d5fd0a(a)syzkaller.appspotmail.com Cc: <stable(a)vger.kernel.org> Link: https://lore.kernel.org/r/00000000000011555605dceaff03@google.com Signed-off-by: Takashi Iwai <tiwai(a)suse.de> --- sound/usb/midi.c | 1 + 1 file changed, 1 insertion(+) diff --git a/sound/usb/midi.c b/sound/usb/midi.c index 2c01649c70f6..7c6ca2b433a5 100644 --- a/sound/usb/midi.c +++ b/sound/usb/midi.c @@ -1194,6 +1194,7 @@ static void snd_usbmidi_output_drain(struct snd_rawmidi_substream *substream) } while (drain_urbs && timeout); finish_wait(&ep->drain_wait, &wait); } + port->active = 0; spin_unlock_irq(&ep->buffer_lock); } -- 2.34.1

1 0

[PATCH] [v3] m68k: coldfire: drop ISA_DMA_API support
by Arnd Bergmann 20 Apr '22

20 Apr '22

From: Arnd Bergmann <arnd(a)arndb.de> After a build regression report, I took a look at possible users of CONFIG_ISA_DMA_API on m68k and found none, which Greg confirmed. The CONFIG_GENERIC_ISA_DMA option in turn is only needed to implement ISA_DMA_API, and is clearly not used on the platforms with ISA support. The CONFIG_ISA support for AMIGA_PCMCIA is probably also unneeded, but this is less clear. Unlike other PCMCIA implementations, this one does not use the drivers/pcmcia subsystem at all and just supports the "apne" network driver. When it was first added, one could use ISA drivers on it as well, but this probably broke at some point. With no reason to keep this, let's just drop the corresponding files and prevent the remaining ISA drivers that use this from getting built. The remaining definitions in asm/dma.h are used for PCI support. Link: https://lore.kernel.org/lkml/9e5ee1c3-ca80-f343-a1f5-66f3dd1c0727@linux-m68… Acked-by: Geert Uytterhoeven <geert(a)linux-m68k.org> Acked-by: Ulf Hansson <ulf.hansson(a)linaro.org> # For MMC Acked-by: Greg Ungerer <gerg(a)linux-m68k.org> Signed-off-by: Arnd Bergmann <arnd(a)arndb.de> --- v3: fix build regression v2: drop GENERIC_ISA_DMA as well, add some background on CONFIG_ISA. --- arch/m68k/Kconfig.bus | 10 - arch/m68k/coldfire/dma.c | 43 ---- arch/m68k/include/asm/dma.h | 483 ------------------------------------ arch/m68k/mm/motorola.c | 1 - drivers/mmc/host/Kconfig | 2 +- sound/core/Makefile | 2 - sound/isa/Kconfig | 2 +- 7 files changed, 2 insertions(+), 541 deletions(-) delete mode 100644 arch/m68k/coldfire/dma.c diff --git a/arch/m68k/Kconfig.bus b/arch/m68k/Kconfig.bus index d1e93a39cd3b..d5c0b2990ee1 100644 --- a/arch/m68k/Kconfig.bus +++ b/arch/m68k/Kconfig.bus @@ -56,16 +56,6 @@ config ATARI_ROM_ISA The only driver currently using this adapter is the EtherNEC driver for RTL8019AS based NE2000 compatible network cards. -config GENERIC_ISA_DMA - def_bool ISA - source "drivers/zorro/Kconfig" endif - -if COLDFIRE - -config ISA_DMA_API - def_bool !M5272 - -endif diff --git a/arch/m68k/coldfire/dma.c b/arch/m68k/coldfire/dma.c deleted file mode 100644 index c3279f7467d7..000000000000 --- a/arch/m68k/coldfire/dma.c +++ /dev/null @@ -1,43 +0,0 @@ -// SPDX-License-Identifier: GPL-2.0 -/***************************************************************************/ - -/* - * dma.c -- Freescale ColdFire DMA support - * - * Copyright (C) 2007, Greg Ungerer (gerg(a)snapgear.com) - */ - -/***************************************************************************/ - -#include <linux/kernel.h> -#include <linux/module.h> -#include <asm/dma.h> -#include <asm/coldfire.h> -#include <asm/mcfsim.h> -#include <asm/mcfdma.h> - -/***************************************************************************/ - -/* - * DMA channel base address table. - */ -unsigned int dma_base_addr[MAX_M68K_DMA_CHANNELS] = { -#ifdef MCFDMA_BASE0 - MCFDMA_BASE0, -#endif -#ifdef MCFDMA_BASE1 - MCFDMA_BASE1, -#endif -#ifdef MCFDMA_BASE2 - MCFDMA_BASE2, -#endif -#ifdef MCFDMA_BASE3 - MCFDMA_BASE3, -#endif -}; -EXPORT_SYMBOL(dma_base_addr); - -unsigned int dma_device_address[MAX_M68K_DMA_CHANNELS]; -EXPORT_SYMBOL(dma_device_address); - -/***************************************************************************/ diff --git a/arch/m68k/include/asm/dma.h b/arch/m68k/include/asm/dma.h index ae2021964e32..f6c5e0dfb4e5 100644 --- a/arch/m68k/include/asm/dma.h +++ b/arch/m68k/include/asm/dma.h @@ -2,493 +2,10 @@ #ifndef _M68K_DMA_H #define _M68K_DMA_H 1 -#ifdef CONFIG_COLDFIRE -/* - * ColdFire DMA Model: - * ColdFire DMA supports two forms of DMA: Single and Dual address. Single - * address mode emits a source address, and expects that the device will either - * pick up the data (DMA READ) or source data (DMA WRITE). This implies that - * the device will place data on the correct byte(s) of the data bus, as the - * memory transactions are always 32 bits. This implies that only 32 bit - * devices will find single mode transfers useful. Dual address DMA mode - * performs two cycles: source read and destination write. ColdFire will - * align the data so that the device will always get the correct bytes, thus - * is useful for 8 and 16 bit devices. This is the mode that is supported - * below. - * - * AUG/22/2000 : added support for 32-bit Dual-Address-Mode (K) 2000 - * Oliver Kamphenkel (O.Kamphenkel(a)tu-bs.de) - * - * AUG/25/2000 : added support for 8, 16 and 32-bit Single-Address-Mode (K)2000 - * Oliver Kamphenkel (O.Kamphenkel(a)tu-bs.de) - * - * APR/18/2002 : added proper support for MCF5272 DMA controller. - * Arthur Shipkowski (art(a)videon-central.com) - */ - -#include <asm/coldfire.h> -#include <asm/mcfsim.h> -#include <asm/mcfdma.h> - -/* - * Set number of channels of DMA on ColdFire for different implementations. - */ -#if defined(CONFIG_M5249) || defined(CONFIG_M5307) || defined(CONFIG_M5407) || \ - defined(CONFIG_M523x) || defined(CONFIG_M527x) || \ - defined(CONFIG_M528x) || defined(CONFIG_M525x) - -#define MAX_M68K_DMA_CHANNELS 4 -#elif defined(CONFIG_M5272) -#define MAX_M68K_DMA_CHANNELS 1 -#elif defined(CONFIG_M53xx) -#define MAX_M68K_DMA_CHANNELS 0 -#else -#define MAX_M68K_DMA_CHANNELS 2 -#endif - -extern unsigned int dma_base_addr[MAX_M68K_DMA_CHANNELS]; -extern unsigned int dma_device_address[MAX_M68K_DMA_CHANNELS]; - -#if !defined(CONFIG_M5272) -#define DMA_MODE_WRITE_BIT 0x01 /* Memory/IO to IO/Memory select */ -#define DMA_MODE_WORD_BIT 0x02 /* 8 or 16 bit transfers */ -#define DMA_MODE_LONG_BIT 0x04 /* or 32 bit transfers */ -#define DMA_MODE_SINGLE_BIT 0x08 /* single-address-mode */ - -/* I/O to memory, 8 bits, mode */ -#define DMA_MODE_READ 0 -/* memory to I/O, 8 bits, mode */ -#define DMA_MODE_WRITE 1 -/* I/O to memory, 16 bits, mode */ -#define DMA_MODE_READ_WORD 2 -/* memory to I/O, 16 bits, mode */ -#define DMA_MODE_WRITE_WORD 3 -/* I/O to memory, 32 bits, mode */ -#define DMA_MODE_READ_LONG 4 -/* memory to I/O, 32 bits, mode */ -#define DMA_MODE_WRITE_LONG 5 -/* I/O to memory, 8 bits, single-address-mode */ -#define DMA_MODE_READ_SINGLE 8 -/* memory to I/O, 8 bits, single-address-mode */ -#define DMA_MODE_WRITE_SINGLE 9 -/* I/O to memory, 16 bits, single-address-mode */ -#define DMA_MODE_READ_WORD_SINGLE 10 -/* memory to I/O, 16 bits, single-address-mode */ -#define DMA_MODE_WRITE_WORD_SINGLE 11 -/* I/O to memory, 32 bits, single-address-mode */ -#define DMA_MODE_READ_LONG_SINGLE 12 -/* memory to I/O, 32 bits, single-address-mode */ -#define DMA_MODE_WRITE_LONG_SINGLE 13 - -#else /* CONFIG_M5272 is defined */ - -/* Source static-address mode */ -#define DMA_MODE_SRC_SA_BIT 0x01 -/* Two bits to select between all four modes */ -#define DMA_MODE_SSIZE_MASK 0x06 -/* Offset to shift bits in */ -#define DMA_MODE_SSIZE_OFF 0x01 -/* Destination static-address mode */ -#define DMA_MODE_DES_SA_BIT 0x10 -/* Two bits to select between all four modes */ -#define DMA_MODE_DSIZE_MASK 0x60 -/* Offset to shift bits in */ -#define DMA_MODE_DSIZE_OFF 0x05 -/* Size modifiers */ -#define DMA_MODE_SIZE_LONG 0x00 -#define DMA_MODE_SIZE_BYTE 0x01 -#define DMA_MODE_SIZE_WORD 0x02 -#define DMA_MODE_SIZE_LINE 0x03 - -/* - * Aliases to help speed quick ports; these may be suboptimal, however. They - * do not include the SINGLE mode modifiers since the MCF5272 does not have a - * mode where the device is in control of its addressing. - */ - -/* I/O to memory, 8 bits, mode */ -#define DMA_MODE_READ ((DMA_MODE_SIZE_BYTE << DMA_MODE_DSIZE_OFF) | (DMA_MODE_SIZE_BYTE << DMA_MODE_SSIZE_OFF) | DMA_SRC_SA_BIT) -/* memory to I/O, 8 bits, mode */ -#define DMA_MODE_WRITE ((DMA_MODE_SIZE_BYTE << DMA_MODE_DSIZE_OFF) | (DMA_MODE_SIZE_BYTE << DMA_MODE_SSIZE_OFF) | DMA_DES_SA_BIT) -/* I/O to memory, 16 bits, mode */ -#define DMA_MODE_READ_WORD ((DMA_MODE_SIZE_WORD << DMA_MODE_DSIZE_OFF) | (DMA_MODE_SIZE_WORD << DMA_MODE_SSIZE_OFF) | DMA_SRC_SA_BIT) -/* memory to I/O, 16 bits, mode */ -#define DMA_MODE_WRITE_WORD ((DMA_MODE_SIZE_WORD << DMA_MODE_DSIZE_OFF) | (DMA_MODE_SIZE_WORD << DMA_MODE_SSIZE_OFF) | DMA_DES_SA_BIT) -/* I/O to memory, 32 bits, mode */ -#define DMA_MODE_READ_LONG ((DMA_MODE_SIZE_LONG << DMA_MODE_DSIZE_OFF) | (DMA_MODE_SIZE_LONG << DMA_MODE_SSIZE_OFF) | DMA_SRC_SA_BIT) -/* memory to I/O, 32 bits, mode */ -#define DMA_MODE_WRITE_LONG ((DMA_MODE_SIZE_LONG << DMA_MODE_DSIZE_OFF) | (DMA_MODE_SIZE_LONG << DMA_MODE_SSIZE_OFF) | DMA_DES_SA_BIT) - -#endif /* !defined(CONFIG_M5272) */ - -#if !defined(CONFIG_M5272) -/* enable/disable a specific DMA channel */ -static __inline__ void enable_dma(unsigned int dmanr) -{ - volatile unsigned short *dmawp; - -#ifdef DMA_DEBUG - printk("enable_dma(dmanr=%d)\n", dmanr); -#endif - - dmawp = (unsigned short *) dma_base_addr[dmanr]; - dmawp[MCFDMA_DCR] |= MCFDMA_DCR_EEXT; -} - -static __inline__ void disable_dma(unsigned int dmanr) -{ - volatile unsigned short *dmawp; - volatile unsigned char *dmapb; - -#ifdef DMA_DEBUG - printk("disable_dma(dmanr=%d)\n", dmanr); -#endif - - dmawp = (unsigned short *) dma_base_addr[dmanr]; - dmapb = (unsigned char *) dma_base_addr[dmanr]; - - /* Turn off external requests, and stop any DMA in progress */ - dmawp[MCFDMA_DCR] &= ~MCFDMA_DCR_EEXT; - dmapb[MCFDMA_DSR] = MCFDMA_DSR_DONE; -} - -/* - * Clear the 'DMA Pointer Flip Flop'. - * Write 0 for LSB/MSB, 1 for MSB/LSB access. - * Use this once to initialize the FF to a known state. - * After that, keep track of it. :-) - * --- In order to do that, the DMA routines below should --- - * --- only be used while interrupts are disabled! --- - * - * This is a NOP for ColdFire. Provide a stub for compatibility. - */ -static __inline__ void clear_dma_ff(unsigned int dmanr) -{ -} - -/* set mode (above) for a specific DMA channel */ -static __inline__ void set_dma_mode(unsigned int dmanr, char mode) -{ - - volatile unsigned char *dmabp; - volatile unsigned short *dmawp; - -#ifdef DMA_DEBUG - printk("set_dma_mode(dmanr=%d,mode=%d)\n", dmanr, mode); -#endif - - dmabp = (unsigned char *) dma_base_addr[dmanr]; - dmawp = (unsigned short *) dma_base_addr[dmanr]; - - /* Clear config errors */ - dmabp[MCFDMA_DSR] = MCFDMA_DSR_DONE; - - /* Set command register */ - dmawp[MCFDMA_DCR] = - MCFDMA_DCR_INT | /* Enable completion irq */ - MCFDMA_DCR_CS | /* Force one xfer per request */ - MCFDMA_DCR_AA | /* Enable auto alignment */ - /* single-address-mode */ - ((mode & DMA_MODE_SINGLE_BIT) ? MCFDMA_DCR_SAA : 0) | - /* sets s_rw (-> r/w) high if Memory to I/0 */ - ((mode & DMA_MODE_WRITE_BIT) ? MCFDMA_DCR_S_RW : 0) | - /* Memory to I/O or I/O to Memory */ - ((mode & DMA_MODE_WRITE_BIT) ? MCFDMA_DCR_SINC : MCFDMA_DCR_DINC) | - /* 32 bit, 16 bit or 8 bit transfers */ - ((mode & DMA_MODE_WORD_BIT) ? MCFDMA_DCR_SSIZE_WORD : - ((mode & DMA_MODE_LONG_BIT) ? MCFDMA_DCR_SSIZE_LONG : - MCFDMA_DCR_SSIZE_BYTE)) | - ((mode & DMA_MODE_WORD_BIT) ? MCFDMA_DCR_DSIZE_WORD : - ((mode & DMA_MODE_LONG_BIT) ? MCFDMA_DCR_DSIZE_LONG : - MCFDMA_DCR_DSIZE_BYTE)); - -#ifdef DEBUG_DMA - printk("%s(%d): dmanr=%d DSR[%x]=%x DCR[%x]=%x\n", __FILE__, __LINE__, - dmanr, (int) &dmabp[MCFDMA_DSR], dmabp[MCFDMA_DSR], - (int) &dmawp[MCFDMA_DCR], dmawp[MCFDMA_DCR]); -#endif -} - -/* Set transfer address for specific DMA channel */ -static __inline__ void set_dma_addr(unsigned int dmanr, unsigned int a) -{ - volatile unsigned short *dmawp; - volatile unsigned int *dmalp; - -#ifdef DMA_DEBUG - printk("set_dma_addr(dmanr=%d,a=%x)\n", dmanr, a); -#endif - - dmawp = (unsigned short *) dma_base_addr[dmanr]; - dmalp = (unsigned int *) dma_base_addr[dmanr]; - - /* Determine which address registers are used for memory/device accesses */ - if (dmawp[MCFDMA_DCR] & MCFDMA_DCR_SINC) { - /* Source incrementing, must be memory */ - dmalp[MCFDMA_SAR] = a; - /* Set dest address, must be device */ - dmalp[MCFDMA_DAR] = dma_device_address[dmanr]; - } else { - /* Destination incrementing, must be memory */ - dmalp[MCFDMA_DAR] = a; - /* Set source address, must be device */ - dmalp[MCFDMA_SAR] = dma_device_address[dmanr]; - } - -#ifdef DEBUG_DMA - printk("%s(%d): dmanr=%d DCR[%x]=%x SAR[%x]=%08x DAR[%x]=%08x\n", - __FILE__, __LINE__, dmanr, (int) &dmawp[MCFDMA_DCR], dmawp[MCFDMA_DCR], - (int) &dmalp[MCFDMA_SAR], dmalp[MCFDMA_SAR], - (int) &dmalp[MCFDMA_DAR], dmalp[MCFDMA_DAR]); -#endif -} - -/* - * Specific for Coldfire - sets device address. - * Should be called after the mode set call, and before set DMA address. - */ -static __inline__ void set_dma_device_addr(unsigned int dmanr, unsigned int a) -{ -#ifdef DMA_DEBUG - printk("set_dma_device_addr(dmanr=%d,a=%x)\n", dmanr, a); -#endif - - dma_device_address[dmanr] = a; -} - -/* - * NOTE 2: "count" represents _bytes_. - */ -static __inline__ void set_dma_count(unsigned int dmanr, unsigned int count) -{ - volatile unsigned short *dmawp; - -#ifdef DMA_DEBUG - printk("set_dma_count(dmanr=%d,count=%d)\n", dmanr, count); -#endif - - dmawp = (unsigned short *) dma_base_addr[dmanr]; - dmawp[MCFDMA_BCR] = (unsigned short)count; -} - -/* - * Get DMA residue count. After a DMA transfer, this - * should return zero. Reading this while a DMA transfer is - * still in progress will return unpredictable results. - * Otherwise, it returns the number of _bytes_ left to transfer. - */ -static __inline__ int get_dma_residue(unsigned int dmanr) -{ - volatile unsigned short *dmawp; - unsigned short count; - -#ifdef DMA_DEBUG - printk("get_dma_residue(dmanr=%d)\n", dmanr); -#endif - - dmawp = (unsigned short *) dma_base_addr[dmanr]; - count = dmawp[MCFDMA_BCR]; - return((int) count); -} -#else /* CONFIG_M5272 is defined */ - -/* - * The MCF5272 DMA controller is very different than the controller defined above - * in terms of register mapping. For instance, with the exception of the 16-bit - * interrupt register (IRQ#85, for reference), all of the registers are 32-bit. - * - * The big difference, however, is the lack of device-requested DMA. All modes - * are dual address transfer, and there is no 'device' setup or direction bit. - * You can DMA between a device and memory, between memory and memory, or even between - * two devices directly, with any combination of incrementing and non-incrementing - * addresses you choose. This puts a crimp in distinguishing between the 'device - * address' set up by set_dma_device_addr. - * - * Therefore, there are two options. One is to use set_dma_addr and set_dma_device_addr, - * which will act exactly as above in -- it will look to see if the source is set to - * autoincrement, and if so it will make the source use the set_dma_addr value and the - * destination the set_dma_device_addr value. Otherwise the source will be set to the - * set_dma_device_addr value and the destination will get the set_dma_addr value. - * - * The other is to use the provided set_dma_src_addr and set_dma_dest_addr functions - * and make it explicit. Depending on what you're doing, one of these two should work - * for you, but don't mix them in the same transfer setup. - */ - -/* enable/disable a specific DMA channel */ -static __inline__ void enable_dma(unsigned int dmanr) -{ - volatile unsigned int *dmalp; - -#ifdef DMA_DEBUG - printk("enable_dma(dmanr=%d)\n", dmanr); -#endif - - dmalp = (unsigned int *) dma_base_addr[dmanr]; - dmalp[MCFDMA_DMR] |= MCFDMA_DMR_EN; -} - -static __inline__ void disable_dma(unsigned int dmanr) -{ - volatile unsigned int *dmalp; - -#ifdef DMA_DEBUG - printk("disable_dma(dmanr=%d)\n", dmanr); -#endif - - dmalp = (unsigned int *) dma_base_addr[dmanr]; - - /* Turn off external requests, and stop any DMA in progress */ - dmalp[MCFDMA_DMR] &= ~MCFDMA_DMR_EN; - dmalp[MCFDMA_DMR] |= MCFDMA_DMR_RESET; -} - -/* - * Clear the 'DMA Pointer Flip Flop'. - * Write 0 for LSB/MSB, 1 for MSB/LSB access. - * Use this once to initialize the FF to a known state. - * After that, keep track of it. :-) - * --- In order to do that, the DMA routines below should --- - * --- only be used while interrupts are disabled! --- - * - * This is a NOP for ColdFire. Provide a stub for compatibility. - */ -static __inline__ void clear_dma_ff(unsigned int dmanr) -{ -} - -/* set mode (above) for a specific DMA channel */ -static __inline__ void set_dma_mode(unsigned int dmanr, char mode) -{ - - volatile unsigned int *dmalp; - volatile unsigned short *dmawp; - -#ifdef DMA_DEBUG - printk("set_dma_mode(dmanr=%d,mode=%d)\n", dmanr, mode); -#endif - dmalp = (unsigned int *) dma_base_addr[dmanr]; - dmawp = (unsigned short *) dma_base_addr[dmanr]; - - /* Clear config errors */ - dmalp[MCFDMA_DMR] |= MCFDMA_DMR_RESET; - - /* Set command register */ - dmalp[MCFDMA_DMR] = - MCFDMA_DMR_RQM_DUAL | /* Mandatory Request Mode setting */ - MCFDMA_DMR_DSTT_SD | /* Set up addressing types; set to supervisor-data. */ - MCFDMA_DMR_SRCT_SD | /* Set up addressing types; set to supervisor-data. */ - /* source static-address-mode */ - ((mode & DMA_MODE_SRC_SA_BIT) ? MCFDMA_DMR_SRCM_SA : MCFDMA_DMR_SRCM_IA) | - /* dest static-address-mode */ - ((mode & DMA_MODE_DES_SA_BIT) ? MCFDMA_DMR_DSTM_SA : MCFDMA_DMR_DSTM_IA) | - /* burst, 32 bit, 16 bit or 8 bit transfers are separately configurable on the MCF5272 */ - (((mode & DMA_MODE_SSIZE_MASK) >> DMA_MODE_SSIZE_OFF) << MCFDMA_DMR_DSTS_OFF) | - (((mode & DMA_MODE_SSIZE_MASK) >> DMA_MODE_SSIZE_OFF) << MCFDMA_DMR_SRCS_OFF); - - dmawp[MCFDMA_DIR] |= MCFDMA_DIR_ASCEN; /* Enable completion interrupts */ - -#ifdef DEBUG_DMA - printk("%s(%d): dmanr=%d DMR[%x]=%x DIR[%x]=%x\n", __FILE__, __LINE__, - dmanr, (int) &dmalp[MCFDMA_DMR], dmalp[MCFDMA_DMR], - (int) &dmawp[MCFDMA_DIR], dmawp[MCFDMA_DIR]); -#endif -} - -/* Set transfer address for specific DMA channel */ -static __inline__ void set_dma_addr(unsigned int dmanr, unsigned int a) -{ - volatile unsigned int *dmalp; - -#ifdef DMA_DEBUG - printk("set_dma_addr(dmanr=%d,a=%x)\n", dmanr, a); -#endif - - dmalp = (unsigned int *) dma_base_addr[dmanr]; - - /* Determine which address registers are used for memory/device accesses */ - if (dmalp[MCFDMA_DMR] & MCFDMA_DMR_SRCM) { - /* Source incrementing, must be memory */ - dmalp[MCFDMA_DSAR] = a; - /* Set dest address, must be device */ - dmalp[MCFDMA_DDAR] = dma_device_address[dmanr]; - } else { - /* Destination incrementing, must be memory */ - dmalp[MCFDMA_DDAR] = a; - /* Set source address, must be device */ - dmalp[MCFDMA_DSAR] = dma_device_address[dmanr]; - } - -#ifdef DEBUG_DMA - printk("%s(%d): dmanr=%d DMR[%x]=%x SAR[%x]=%08x DAR[%x]=%08x\n", - __FILE__, __LINE__, dmanr, (int) &dmalp[MCFDMA_DMR], dmalp[MCFDMA_DMR], - (int) &dmalp[MCFDMA_DSAR], dmalp[MCFDMA_DSAR], - (int) &dmalp[MCFDMA_DDAR], dmalp[MCFDMA_DDAR]); -#endif -} - -/* - * Specific for Coldfire - sets device address. - * Should be called after the mode set call, and before set DMA address. - */ -static __inline__ void set_dma_device_addr(unsigned int dmanr, unsigned int a) -{ -#ifdef DMA_DEBUG - printk("set_dma_device_addr(dmanr=%d,a=%x)\n", dmanr, a); -#endif - - dma_device_address[dmanr] = a; -} - -/* - * NOTE 2: "count" represents _bytes_. - * - * NOTE 3: While a 32-bit register, "count" is only a maximum 24-bit value. - */ -static __inline__ void set_dma_count(unsigned int dmanr, unsigned int count) -{ - volatile unsigned int *dmalp; - -#ifdef DMA_DEBUG - printk("set_dma_count(dmanr=%d,count=%d)\n", dmanr, count); -#endif - - dmalp = (unsigned int *) dma_base_addr[dmanr]; - dmalp[MCFDMA_DBCR] = count; -} - -/* - * Get DMA residue count. After a DMA transfer, this - * should return zero. Reading this while a DMA transfer is - * still in progress will return unpredictable results. - * Otherwise, it returns the number of _bytes_ left to transfer. - */ -static __inline__ int get_dma_residue(unsigned int dmanr) -{ - volatile unsigned int *dmalp; - unsigned int count; - -#ifdef DMA_DEBUG - printk("get_dma_residue(dmanr=%d)\n", dmanr); -#endif - - dmalp = (unsigned int *) dma_base_addr[dmanr]; - count = dmalp[MCFDMA_DBCR]; - return(count); -} - -#endif /* !defined(CONFIG_M5272) */ -#endif /* CONFIG_COLDFIRE */ - /* it's useless on the m68k, but unfortunately needed by the new bootmem allocator (but this should do it for this) */ #define MAX_DMA_ADDRESS PAGE_OFFSET -#define MAX_DMA_CHANNELS 8 - -extern int request_dma(unsigned int dmanr, const char * device_id); /* reserve a DMA channel */ -extern void free_dma(unsigned int dmanr); /* release it again */ - #ifdef CONFIG_PCI extern int isa_dma_bridge_buggy; #else diff --git a/arch/m68k/mm/motorola.c b/arch/m68k/mm/motorola.c index ecbe948f4c1a..df7f797c908a 100644 --- a/arch/m68k/mm/motorola.c +++ b/arch/m68k/mm/motorola.c @@ -27,7 +27,6 @@ #include <asm/pgalloc.h> #include <asm/machdep.h> #include <asm/io.h> -#include <asm/dma.h> #ifdef CONFIG_ATARI #include <asm/atari_stram.h> #endif diff --git a/drivers/mmc/host/Kconfig b/drivers/mmc/host/Kconfig index af6c3c329076..d6144978e32d 100644 --- a/drivers/mmc/host/Kconfig +++ b/drivers/mmc/host/Kconfig @@ -508,7 +508,7 @@ config MMC_OMAP_HS config MMC_WBSD tristate "Winbond W83L51xD SD/MMC Card Interface support" - depends on ISA_DMA_API && !M68K + depends on ISA_DMA_API help This selects the Winbond(R) W83L51xD Secure digital and Multimedia card Interface. diff --git a/sound/core/Makefile b/sound/core/Makefile index 350d704ced98..2762f03d9b7b 100644 --- a/sound/core/Makefile +++ b/sound/core/Makefile @@ -9,9 +9,7 @@ ifneq ($(CONFIG_SND_PROC_FS),) snd-y += info.o snd-$(CONFIG_SND_OSSEMUL) += info_oss.o endif -ifneq ($(CONFIG_M68K),y) snd-$(CONFIG_ISA_DMA_API) += isadma.o -endif snd-$(CONFIG_SND_OSSEMUL) += sound_oss.o snd-$(CONFIG_SND_VMASTER) += vmaster.o snd-$(CONFIG_SND_JACK) += ctljack.o jack.o diff --git a/sound/isa/Kconfig b/sound/isa/Kconfig index 570b88e0b201..6ffa48dd5983 100644 --- a/sound/isa/Kconfig +++ b/sound/isa/Kconfig @@ -22,7 +22,7 @@ config SND_SB16_DSP menuconfig SND_ISA bool "ISA sound devices" depends on ISA || COMPILE_TEST - depends on ISA_DMA_API && !M68K + depends on ISA_DMA_API default y help Support for sound devices connected via the ISA bus. -- 2.29.2

2 1