Slightly brushing up and throw the old dust away from my ancient writing-an-alsa-driver document. The contents aren't changed so much but the obsoleted parts are dropped.
Also, remove the date and the version number. It's useless.
Signed-off-by: Takashi Iwai tiwai@suse.de --- .../kernel-api/writing-an-alsa-driver.rst | 307 +++++++++--------- 1 file changed, 149 insertions(+), 158 deletions(-)
diff --git a/Documentation/sound/kernel-api/writing-an-alsa-driver.rst b/Documentation/sound/kernel-api/writing-an-alsa-driver.rst index a0b268466cb1..b60a26807420 100644 --- a/Documentation/sound/kernel-api/writing-an-alsa-driver.rst +++ b/Documentation/sound/kernel-api/writing-an-alsa-driver.rst @@ -3,8 +3,6 @@ Writing an ALSA Driver ======================
:Author: Takashi Iwai tiwai@suse.de -:Date: Oct 15, 2007 -:Edition: 0.3.7
Preface ======= @@ -21,11 +19,6 @@ explain the general topic of linux kernel coding and doesn't cover low-level driver implementation details. It only describes the standard way to write a PCI sound driver on ALSA.
-If you are already familiar with the older ALSA ver.0.5.x API, you can -check the drivers such as ``sound/pci/es1938.c`` or -``sound/pci/maestro3.c`` which have also almost the same code-base in -the ALSA 0.5.x tree, so you can compare the differences. - This document is still a draft version. Any feedback and corrections, please!!
@@ -35,24 +28,7 @@ File Tree Structure General -------
-The ALSA drivers are provided in two ways. - -One is the trees provided as a tarball or via cvs from the ALSA's ftp -site, and another is the 2.6 (or later) Linux kernel tree. To -synchronize both, the ALSA driver tree is split into two different -trees: alsa-kernel and alsa-driver. The former contains purely the -source code for the Linux 2.6 (or later) tree. This tree is designed -only for compilation on 2.6 or later environment. The latter, -alsa-driver, contains many subtle files for compiling ALSA drivers -outside of the Linux kernel tree, wrapper functions for older 2.2 and -2.4 kernels, to adapt the latest kernel API, and additional drivers -which are still in development or in tests. The drivers in alsa-driver -tree will be moved to alsa-kernel (and eventually to the 2.6 kernel -tree) when they are finished and confirmed to work fine. - -The file tree structure of ALSA driver is depicted below. Both -alsa-kernel and alsa-driver have almost the same file structure, except -for “core” directory. It's named as “acore” in alsa-driver tree. +The file tree structure of ALSA driver is depicted below.
::
@@ -61,14 +37,11 @@ for “core” directory. It's named as “acore” in alsa-driver tree. /oss /seq /oss - /instr - /ioctl32 /include /drivers /mpu401 /opl3 /i2c - /l3 /synth /emux /pci @@ -80,6 +53,7 @@ for “core” directory. It's named as “acore” in alsa-driver tree. /sparc /usb /pcmcia /(cards) + /soc /oss
@@ -99,13 +73,6 @@ directory. The rawmidi OSS emulation is included in the ALSA rawmidi code since it's quite small. The sequencer code is stored in ``core/seq/oss`` directory (see `below <#core-seq-oss>`__).
-core/ioctl32 -~~~~~~~~~~~~ - -This directory contains the 32bit-ioctl wrappers for 64bit architectures -such like x86-64, ppc64 and sparc64. For 32bit and alpha architectures, -these are not compiled. - core/seq ~~~~~~~~
@@ -119,11 +86,6 @@ core/seq/oss
This contains the OSS sequencer emulation codes.
-core/seq/instr -~~~~~~~~~~~~~~ - -This directory contains the modules for the sequencer instrument layer. - include directory -----------------
@@ -161,11 +123,6 @@ Although there is a standard i2c layer on Linux, ALSA has its own i2c code for some cards, because the soundcard needs only a simple operation and the standard i2c API is too complicated for such a purpose.
-i2c/l3 -~~~~~~ - -This is a sub-directory for ARM L3 i2c. - synth directory ---------------
@@ -209,11 +166,19 @@ The PCMCIA, especially PCCard drivers will go here. CardBus drivers will be in the pci directory, because their API is identical to that of standard PCI cards.
+soc directory +------------- + +This directory contains the codes for ASoC (ALSA System on Chip) +layer including ASoC core, codec and machine drivers. + oss directory -------------
-The OSS/Lite source files are stored here in Linux 2.6 (or later) tree. -In the ALSA driver tarball, this directory is empty, of course :) +Here contains OSS/Lite codes. +All codes have been deprecated except for dmasound on m68k as of +writing this. +
Basic Flow for PCI Drivers ========================== @@ -352,10 +317,8 @@ to details explained in the following section.
/* (3) */ err = snd_mychip_create(card, pci, &chip); - if (err < 0) { - snd_card_free(card); - return err; - } + if (err < 0) + goto error;
/* (4) */ strcpy(card->driver, "My Chip"); @@ -368,22 +331,23 @@ to details explained in the following section.
/* (6) */ err = snd_card_register(card); - if (err < 0) { - snd_card_free(card); - return err; - } + if (err < 0) + goto error;
/* (7) */ pci_set_drvdata(pci, card); dev++; return 0; + + error: + snd_card_free(card); + return err; }
/* destructor -- see the "Destructor" sub-section */ static void snd_mychip_remove(struct pci_dev *pci) { snd_card_free(pci_get_drvdata(pci)); - pci_set_drvdata(pci, NULL); }
@@ -445,14 +409,26 @@ In this part, the PCI resources are allocated. struct mychip *chip; .... err = snd_mychip_create(card, pci, &chip); - if (err < 0) { - snd_card_free(card); - return err; - } + if (err < 0) + goto error;
The details will be explained in the section `PCI Resource Management`_.
+When something goes wrong, the probe function needs to deal with the +error. In this example, we have a single error handling path placed +at the end of the function. + +:: + + error: + snd_card_free(card); + return err; + +Since each component can be properly freed, the single +:c:func:`snd_card_free()` call should suffice in most cases. + + 4) Set the driver ID and name strings. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
@@ -486,10 +462,8 @@ too. ::
err = snd_card_register(card); - if (err < 0) { - snd_card_free(card); - return err; - } + if (err < 0) + goto error;
Will be explained in the section `Management of Cards and Components`_, too. @@ -513,14 +487,13 @@ The destructor, remove callback, simply releases the card instance. Then the ALSA middle layer will release all the attached components automatically.
-It would be typically like the following: +It would be typically just :c:func:`calling snd_card_free()`:
::
static void snd_mychip_remove(struct pci_dev *pci) { snd_card_free(pci_get_drvdata(pci)); - pci_set_drvdata(pci, NULL); }
@@ -546,7 +519,7 @@ in the source file. If the code is split into several files, the files without module options don't need them.
In addition to these headers, you'll need ``<linux/interrupt.h>`` for -interrupt handling, and ``<asm/io.h>`` for I/O access. If you use the +interrupt handling, and ``<linux/io.h>`` for I/O access. If you use the :c:func:`mdelay()` or :c:func:`udelay()` functions, you'll need to include ``<linux/delay.h>`` too.
@@ -720,6 +693,13 @@ function, which will call the real destructor.
where :c:func:`snd_mychip_free()` is the real destructor.
+The demerit of this method is the obviously more amount of codes. +The merit is, however, you can trigger the own callback at registering +and disconnecting the card via setting in snd_device_ops. +About the registering and disconnecting the card, see the subsections +below. + + Registration and Release ------------------------
@@ -905,10 +885,8 @@ Resource Allocation -------------------
The allocation of I/O ports and irqs is done via standard kernel -functions. Unlike ALSA ver.0.5.x., there are no helpers for that. And -these resources must be released in the destructor function (see below). -Also, on ALSA 0.9.x, you don't need to allocate (pseudo-)DMA for PCI -like in ALSA 0.5.x. +functions. These resources must be released in the destructor +function (see below).
Now assume that the PCI device has an I/O port with 8 bytes and an interrupt. Then :c:type:`struct mychip <mychip>` will have the @@ -1064,7 +1042,8 @@ and the allocation would be like below:
::
- if ((err = pci_request_regions(pci, "My Chip")) < 0) { + err = pci_request_regions(pci, "My Chip"); + if (err < 0) { kfree(chip); return err; } @@ -1086,6 +1065,21 @@ and the corresponding destructor would be: .... }
+Of course, a modern way with :c:func:`pci_iomap()` will make things a +bit easier, too. + +:: + + err = pci_request_regions(pci, "My Chip"); + if (err < 0) { + kfree(chip); + return err; + } + chip->iobase_virt = pci_iomap(pci, 0, 0); + +which is paired with :c:func:`pci_iounmap()` at destructor. + + PCI Entries -----------
@@ -1154,13 +1148,6 @@ And at last, the module entries: Note that these module entries are tagged with ``__init`` and ``__exit`` prefixes.
-Oh, one thing was forgotten. If you have no exported symbols, you need -to declare it in 2.2 or 2.4 kernels (it's not necessary in 2.6 kernels). - -:: - - EXPORT_NO_SYMBOLS; - That's all!
PCM Interface @@ -2113,6 +2100,16 @@ non-contiguous buffers. The mmap calls this callback to get the page address. Some examples will be explained in the later section `Buffer and Memory Management`_, too.
+mmap calllback +~~~~~~~~~~~~~~ + +This is another optional callback for controlling mmap behavior. +Once when defined, PCM core calls this callback when a page is +memory-mapped instead of dealing via the standard helper. +If you need special handling (due to some architecture or +device-specific issues), implement everything here as you like. + + PCM Interrupt Handler ---------------------
@@ -2370,6 +2367,27 @@ to define the inverse rule: hw_rule_format_by_channels, NULL, SNDRV_PCM_HW_PARAM_CHANNELS, -1);
+One typical usage of the hw constraints is to align the buffer size +with the period size. As default, ALSA PCM core doesn't enforce the +buffer size to be aligned with the period size. For example, it'd be +possible to have a combination like 256 period bytes with 999 buffer +bytes. + +Many device chips, however, require the buffer to be a multiple of +periods. In such a case, call +:c:func:`snd_pcm_hw_constraint_integer()` for +``SNDRV_PCM_HW_PARAM_PERIODS``. + +:: + + snd_pcm_hw_constraint_integer(substream->runtime, + SNDRV_PCM_HW_PARAM_PERIODS); + +This assures that the number of periods is integer, hence the buffer +size is aligned with the period size. + +The hw constraint is a very much powerful mechanism to define the +preferred PCM configuration, and there are relevant helpers. I won't give more details here, rather I would like to say, “Luke, use the source.”
@@ -3712,7 +3730,14 @@ example, for an intermediate buffer. Since the allocated pages are not contiguous, you need to set the ``page`` callback to obtain the physical address at every offset.
-The implementation of ``page`` callback would be like this: +The easiest way to achieve it would be to use +:c:func:`snd_pcm_lib_alloc_vmalloc_buffer()` for allocating the buffer +via :c:func:`vmalloc()`, and set :c:func:`snd_pcm_sgbuf_ops_page()` to +the ``page`` callback. At release, you need to call +:c:func:`snd_pcm_lib_free_vmalloc_buffer()`. + +If you want to implementation the ``page`` manually, it would be like +this:
::
@@ -3848,7 +3873,9 @@ Power Management
If the chip is supposed to work with suspend/resume functions, you need to add power-management code to the driver. The additional code for -power-management should be ifdef-ed with ``CONFIG_PM``. +power-management should be ifdef-ed with ``CONFIG_PM``, or annotated +with __maybe_unused attribute; otherwise the compiler will complain +you.
If the driver *fully* supports suspend/resume that is, the device can be properly resumed to its state when suspend was called, you can set the @@ -3879,18 +3906,16 @@ the case of PCI drivers, the callbacks look like below:
::
- #ifdef CONFIG_PM - static int snd_my_suspend(struct pci_dev *pci, pm_message_t state) + static int __maybe_unused snd_my_suspend(struct device *dev) { .... /* do things for suspend */ return 0; } - static int snd_my_resume(struct pci_dev *pci) + static int __maybe_unused snd_my_resume(struct device *dev) { .... /* do things for suspend */ return 0; } - #endif
The scheme of the real suspend job is as follows.
@@ -3909,18 +3934,14 @@ The scheme of the real suspend job is as follows.
6. Stop the hardware if necessary.
-7. Disable the PCI device by calling - :c:func:`pci_disable_device()`. Then, call - :c:func:`pci_save_state()` at last. - A typical code would be like:
::
- static int mychip_suspend(struct pci_dev *pci, pm_message_t state) + static int __maybe_unused mychip_suspend(struct device *dev) { /* (1) */ - struct snd_card *card = pci_get_drvdata(pci); + struct snd_card *card = dev_get_drvdata(dev); struct mychip *chip = card->private_data; /* (2) */ snd_power_change_state(card, SNDRV_CTL_POWER_D3hot); @@ -3932,9 +3953,6 @@ A typical code would be like: snd_mychip_save_registers(chip); /* (6) */ snd_mychip_stop_hardware(chip); - /* (7) */ - pci_disable_device(pci); - pci_save_state(pci); return 0; }
@@ -3943,44 +3961,35 @@ The scheme of the real resume job is as follows.
1. Retrieve the card and the chip data.
-2. Set up PCI. First, call :c:func:`pci_restore_state()`. Then - enable the pci device again by calling - :c:func:`pci_enable_device()`. Call - :c:func:`pci_set_master()` if necessary, too. +2. Re-initialize the chip.
-3. Re-initialize the chip. +3. Restore the saved registers if necessary.
-4. Restore the saved registers if necessary. +4. Resume the mixer, e.g. calling :c:func:`snd_ac97_resume()`.
-5. Resume the mixer, e.g. calling :c:func:`snd_ac97_resume()`. +5. Restart the hardware (if any).
-6. Restart the hardware (if any). - -7. Call :c:func:`snd_power_change_state()` with +6. Call :c:func:`snd_power_change_state()` with ``SNDRV_CTL_POWER_D0`` to notify the processes.
A typical code would be like:
::
- static int mychip_resume(struct pci_dev *pci) + static int __maybe_unused mychip_resume(struct pci_dev *pci) { /* (1) */ - struct snd_card *card = pci_get_drvdata(pci); + struct snd_card *card = dev_get_drvdata(dev); struct mychip *chip = card->private_data; /* (2) */ - pci_restore_state(pci); - pci_enable_device(pci); - pci_set_master(pci); - /* (3) */ snd_mychip_reinit_chip(chip); - /* (4) */ + /* (3) */ snd_mychip_restore_registers(chip); - /* (5) */ + /* (4) */ snd_ac97_resume(chip->ac97); - /* (6) */ + /* (5) */ snd_mychip_restart_chip(chip); - /* (7) */ + /* (6) */ snd_power_change_state(card, SNDRV_CTL_POWER_D0); return 0; } @@ -4046,15 +4055,14 @@ And next, set suspend/resume callbacks to the pci_driver.
::
+ static SIMPLE_DEV_PM_OPS(snd_my_pm_ops, mychip_suspend, mychip_resume); + static struct pci_driver driver = { .name = KBUILD_MODNAME, .id_table = snd_my_ids, .probe = snd_my_probe, .remove = snd_my_remove, - #ifdef CONFIG_PM - .suspend = snd_my_suspend, - .resume = snd_my_resume, - #endif + .driver.pm = &snd_my_pm_ops, };
Module Parameters @@ -4078,7 +4086,7 @@ variables, instead. ``enable`` option is not always necessary in this case, but it would be better to have a dummy option for compatibility.
The module parameters must be declared with the standard -``module_param()()``, ``module_param_array()()`` and +``module_param()``, ``module_param_array()`` and :c:func:`MODULE_PARM_DESC()` macros.
The typical coding would be like below: @@ -4094,15 +4102,14 @@ The typical coding would be like below: module_param_array(enable, bool, NULL, 0444); MODULE_PARM_DESC(enable, "Enable " CARD_NAME " soundcard.");
-Also, don't forget to define the module description, classes, license -and devices. Especially, the recent modprobe requires to define the +Also, don't forget to define the module description and the license. +Especially, the recent modprobe requires to define the module license as GPL, etc., otherwise the system is shown as “tainted”.
::
- MODULE_DESCRIPTION("My Chip"); + MODULE_DESCRIPTION("Sound driver for My Chip"); MODULE_LICENSE("GPL"); - MODULE_SUPPORTED_DEVICE("{{Vendor,My Chip Name}}");
How To Put Your Driver Into ALSA Tree @@ -4117,21 +4124,17 @@ a question now: how to put my own driver into the ALSA driver tree? Here
Suppose that you create a new PCI driver for the card “xyz”. The card module name would be snd-xyz. The new driver is usually put into the -alsa-driver tree, ``alsa-driver/pci`` directory in the case of PCI -cards. Then the driver is evaluated, audited and tested by developers -and users. After a certain time, the driver will go to the alsa-kernel -tree (to the corresponding directory, such as ``alsa-kernel/pci``) and -eventually will be integrated into the Linux 2.6 tree (the directory -would be ``linux/sound/pci``). +alsa-driver tree, ``sound/pci`` directory in the case of PCI +cards.
In the following sections, the driver code is supposed to be put into -alsa-driver tree. The two cases are covered: a driver consisting of a +Linux kernel tree. The two cases are covered: a driver consisting of a single source file and one consisting of several source files.
Driver with A Single Source File --------------------------------
-1. Modify alsa-driver/pci/Makefile +1. Modify sound/pci/Makefile
Suppose you have a file xyz.c. Add the following two lines
@@ -4160,52 +4163,43 @@ Driver with A Single Source File
For the details of Kconfig script, refer to the kbuild documentation.
-3. Run cvscompile script to re-generate the configure script and build - the whole stuff again. - Drivers with Several Source Files ---------------------------------
Suppose that the driver snd-xyz have several source files. They are -located in the new subdirectory, pci/xyz. +located in the new subdirectory, sound/pci/xyz.
-1. Add a new directory (``xyz``) in ``alsa-driver/pci/Makefile`` as - below +1. Add a new directory (``sound/pci/xyz``) in ``sound/pci/Makefile`` + as below
::
- obj-$(CONFIG_SND) += xyz/ + obj-$(CONFIG_SND) += sound/pci/xyz/
-2. Under the directory ``xyz``, create a Makefile +2. Under the directory ``sound/pci/xyz``, create a Makefile
::
- ifndef SND_TOPDIR - SND_TOPDIR=../.. - endif - - include $(SND_TOPDIR)/toplevel.config - include $(SND_TOPDIR)/Makefile.conf - snd-xyz-objs := xyz.o abc.o def.o - obj-$(CONFIG_SND_XYZ) += snd-xyz.o
- include $(SND_TOPDIR)/Rules.make - 3. Create the Kconfig entry
This procedure is as same as in the last section.
-4. Run cvscompile script to re-generate the configure script and build - the whole stuff again.
Useful Functions ================
:c:func:`snd_printk()` and friends ---------------------------------------- +---------------------------------- + +.. note:: This subsection describes a few helper functions for +decorating a bit more on the standard :c:func:`printk()` & co. +However, in general, the use of such helpers is no longer recommended. +If possible, try to stick with the standard functions like +:c:func:`dev_err()` or :c:func:`pr_err()`.
ALSA provides a verbose version of the :c:func:`printk()` function. If a kernel config ``CONFIG_SND_VERBOSE_PRINTK`` is set, this function @@ -4221,13 +4215,10 @@ just like :c:func:`snd_printk()`. If the ALSA is compiled without the debugging flag, it's ignored.
:c:func:`snd_printdd()` is compiled in only when -``CONFIG_SND_DEBUG_VERBOSE`` is set. Please note that -``CONFIG_SND_DEBUG_VERBOSE`` is not set as default even if you configure -the alsa-driver with ``--with-debug=full`` option. You need to give -explicitly ``--with-debug=detect`` option instead. +``CONFIG_SND_DEBUG_VERBOSE`` is set.
:c:func:`snd_BUG()` ------------------------- +-------------------
It shows the ``BUG?`` message and stack trace as well as :c:func:`snd_BUG_ON()` at the point. It's useful to show that a @@ -4236,7 +4227,7 @@ fatal error happens there. When no debug flag is set, this macro is ignored.
:c:func:`snd_BUG_ON()` ----------------------------- +----------------------
:c:func:`snd_BUG_ON()` macro is similar with :c:func:`WARN_ON()` macro. For example, snd_BUG_ON(!pointer); or