The "dm" command under CONFIG_CMD_DM should live under cmd/ rather than
test/dm/ so move it.
Signed-off-by: NTom Rini <trini@konsulko.com>

The commit below incorrectly fixed hexport_r();
 >	size = totlen + 1;
One extra byte is necessary to NULL-terminate a whole buffer, "resp."

Fixes: f1b20acb ("hashtable: Fix length calculation in hexport_r")
Signed-off-by: NAKASHI Takahiro <takahiro.akashi@linaro.org>

If OP-TEE core is compiled with support of REE FS and RPMB
at the same time (CFG_RPMB_FS ?= y; CFG_RPMB_FS ?= y), and persistent
storage API is used with TEE_STORAGE_PRIVATE storage id, it will
lead to TA panic.

E/TC:? 0 TA panicked with code 0xffff0009
.....
E/TC:? 0 Call stack:
E/TC:? 0  0x000000004002f2f8 TEE_OpenPersistentObject at
lib/libutee/tee_api_objects.c:422

In this particular case TEE_ERROR_STORAGE_NOT_AVAILABLE is more suitable
than TEE_ERROR_NOT_IMPLEMENTED, as it provides to a TA a possibility
to handle this error code [1].

>From GPD TEE Internal Core specification [2]:
TEE_ERROR_STORAGE_NOT_AVAILABLE - if the persistent object is stored in a
storage area which is currently inaccessible. It may be associated with
the device but unplugged, busy, or inaccessible for some other reason.

[1]: https://github.com/OP-TEE/optee_os/blob/94db01ef448d1e552161c2d861d57a5f8bda0cc0/lib/libutee/tee_api_objects.c#L419
[2]: https://globalplatform.org/wp-content/uploads/2018/06/GPD_TEE_Internal_Core_API_Specification_v1.1.2.50_PublicReview.pdfSigned-off-by: NIgor Opaniuk <igor.opaniuk@linaro.org>
Reviewed-by: NJens Wiklander <jens.wiklander@linaro.org>

Complete conversion of sound to driver model

At present the 'beep' sound generates a waveform for only one channel even
if two are being used. This means that the beep is twice the frequency it
should be. Correct this by making it a parameter.

The fix in a previous commit was correct for sandbox but not for other
boards.

Fixes: 03f11e87 ("sound: Correct data output in sound_create_square_wave()")
Signed-off-by: NSimon Glass <sjg@chromium.org>

At present these parameters are hard-coded in the sdl interface code.
Allow them to be specified by the driver instead.
Signed-off-by: NSimon Glass <sjg@chromium.org>

Return a valid error code instead of -1.
Signed-off-by: NSimon Glass <sjg@chromium.org>

Fix a few files whos license headers were not converted to SPDX.
Signed-off-by: NSimon Glass <sjg@chromium.org>

All users of sound are converted to use driver model. Drop the old code
and the CONFIG_DM_SOUND option.
Signed-off-by: NSimon Glass <sjg@chromium.org>

This option is not needed anymore for all exynos boards except arndale.
Update the config.
Signed-off-by: NSimon Glass <sjg@chromium.org>

Update sandbox's device tree and config to use driver model for sound. Use
the double buffer for sound output so that we don't need to wait for the
sound to complete before returning.
Signed-off-by: NSimon Glass <sjg@chromium.org>

Update snow's device tree and config to use driver model for sound. Also
update the others as best we can.

Spring does not appear to have audio support in the kernel. The smdk5250
and smdk5420 boards use a wolfson codec which I cannot test with. So the
only boards that is tested and known to work are snow, pit and pi.
Signed-off-by: NSimon Glass <sjg@chromium.org>

Add support for this new codec which is used by pit.
Signed-off-by: NSimon Glass <sjg@chromium.org>

The register-access code is useful for any maxim codec. Move it out into
its own file.
Signed-off-by: NSimon Glass <sjg@chromium.org>

Allow setting the i2s pinmux correctly on exyno5420 so that i2c can be
used on that SoC. Also rename EXYNOS_AUDSS to something consistent with
other naming.
Signed-off-by: NSimon Glass <sjg@chromium.org>

Add a convenience function for any Exynos 542x chip.
Signed-off-by: NSimon Glass <sjg@chromium.org>

Move the existing hardware drivers over to use driver model.
Signed-off-by: NSimon Glass <sjg@chromium.org>

The current ID enums start from 1 but there does not seem to be any reason
that they cannot start with 0. Adjust the code to avoid the +1 in
codec_init().
Signed-off-by: NSimon Glass <sjg@chromium.org>

The current dev_read...() functions use s32 and u32 which are convenient
for device tree but not so useful for normal code, which often wants to
use normal integers for values.

Add a helper which supports returning an unsigned int. Also add signed
versions of the unsigned readers.
Signed-off-by: NSimon Glass <sjg@chromium.org>

The sound driver pulls together the audio codec and i2s drivers in order
to actually make sounds. It supports setup() and play() methods. The
sound_find_codec_i2s() function allows locating the linked codec and i2s
devices. They can be referred to from uclass-private data.

Add a uclass and a test for sound.
Signed-off-by: NSimon Glass <sjg@chromium.org>

At present we use a single buffer for sound which means we cannot be
playing one sound while queueing up the next. This wouldn't matter except
that a long sound (more than a second) has to be created as a single
buffer, thus using a lot of memory. To better mimic what real sound
drivers do, add support for double buffering in sandbox.
Signed-off-by: NSimon Glass <sjg@chromium.org>

The i2s bus is commonly used with audio codecs. It provides a way to
stream digital data sychronously in both directions. U-Boot only supports
audio output, so this uclass is very simple, with a single tx_data()
method.

Add a uclass and a test for i2s.
Signed-off-by: NSimon Glass <sjg@chromium.org>

An audio codec provides a way to convert digital data to sound and vice
versa. Add a simple uclass which just supports setting the parameters for
the codec.
Signed-off-by: NSimon Glass <sjg@chromium.org>

This structure contains information that is likely needed by any i2s
driver so it seems useful to attach it to the (forthcoming) i2c uclass.
For now, just rename it.
Signed-off-by: NSimon Glass <sjg@chromium.org>

The U-Boot sound system provides basic support for beeping. At present it
does not use driver model, but it needs to be converted. Add an option to
enable driver model for sound. For now it is not connected to anything.
Future work will add drivers which use this option. It will then be
removed once everything is converted.
Signed-off-by: NSimon Glass <sjg@chromium.org>

For snow the codec is at address 0x11 on the i2c bus, in 7-bit format.
The device tree and code are in 8-bit format (i.e. shifted left one bit).
Fix both. Fix pit in a similar way.
Signed-off-by: NSimon Glass <sjg@chromium.org>

Tidy up the ordering of header files in the sounds files.
Signed-off-by: NSimon Glass <sjg@chromium.org>

This function has only one line in it. Drop it.
Signed-off-by: NSimon Glass <sjg@chromium.org>

These are only used in two functions so can be made local.

Also change the first argument of max98095_do_init() to suit.
Signed-off-by: NSimon Glass <sjg@chromium.org>

These are only used in two functions so can be made local.
Signed-off-by: NSimon Glass <sjg@chromium.org>

We can put this in the private structure and avoid a global.
Signed-off-by: NSimon Glass <sjg@chromium.org>

We can put this in the private structure and avoid a global.
Signed-off-by: NSimon Glass <sjg@chromium.org>

With driver model we want to do a minimal probe when the device is probed
and then set up the codec interface later when a sound is actully played.

Split this setup code out into its own function to help with this.
Signed-off-by: NSimon Glass <sjg@chromium.org>

With driver model we want to do a minimal probe when the device is probed
and then set up the codec interface later when a sound is actully played.

Split this setup code out into its own function to help with this.
Signed-off-by: NSimon Glass <sjg@chromium.org>

With driver model we cannot pass in the global struct, but instead want
to pass in the driver-private data. Split some of the code out of
wm8994_init() to handle this.
Signed-off-by: NSimon Glass <sjg@chromium.org>

At present the driver-private data is global. To allow this code to be
used with driver model, change it to pass the data down to each function.
Use the name 'priv' consistently throughout.

Also rename max98095_update_bits() to max98095_bic_or() which is more
descriptive and shorter, thus breaking fewer lines with the parameter
addition.
Signed-off-by: NSimon Glass <sjg@chromium.org>

At present the driver-private data is global. To allow this code to be
used with driver model, change it to pass the data down to each function.
Use the name 'priv' consistently throughout.

Also rename wm8994_update_bits() to wm8994_bic_or() which is more
descriptive and shorter, thus breaking fewer lines with the parameter
addition.
Signed-off-by: NSimon Glass <sjg@chromium.org>

This field is not really used. It is always set to a known value. Drop it
to simplify the code.
Signed-off-by: NSimon Glass <sjg@chromium.org>

This struct is only used by the Samsung I2C driver and should move into
that driver. For now, rename it so it is clear that is driver-private
info.
Signed-off-by: NSimon Glass <sjg@chromium.org>

CONFIG_OF_CONTROL is enabled for all boards that use sound, so remove the
dead code.
Signed-off-by: NSimon Glass <sjg@chromium.org>