Re: [PATCH resend v5 2/3] ASoC: cygnus: Add Cygnus audio DAI driver

[Mark Brown <broonie@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 2596 bytes --]

On Tue, Mar 29, 2016 at 11:46:31AM -0700, Simran Rai wrote:

A few issues here, a lot of them are stylistic though there's what look
like a couple of small bugs here too.

> +static int cygnus_ssp_set_clocks(struct cygnus_aio_port *aio,
> +			struct cygnus_audio *cygaud)
> +{
> +	u32 value, i = 0;
> +	u32 mask = 0xf;
> +	u32 sclk;
> +	bool found = false;
> +	const struct _ssp_clk_coeff *p_entry = NULL;
> +
> +	if ((!aio->lrclk) || (!aio->bit_per_frame)) {
> +		dev_err(aio->cygaud->dev, "First set up port through hw_params()\n");
> +		return -EINVAL;
> +	}

This function is only ever called from one site in hw_prams().  What is
this defending against?  A check like this seems very worrying, if it
ever goes off that seems to indicate either something is seriously wrong
or we should be recording something then coming back and trying again
later.

> +	for (i = 0; i < ARRAY_SIZE(ssp_clk_coeff); i++) {
> +		p_entry = &ssp_clk_coeff[i];
> +		if ((p_entry->rate == aio->lrclk) &&
> +				(p_entry->sclk_rate == aio->bit_per_frame) &&
> +				(p_entry->mclk == aio->mclk)) {

Why the strange indentation here?

> +	/* Set sclk rate */
> +	if (aio->port_type == PORT_TDM) {

switch statment here for extensibility.

> +		/* Configure channels as mono or stereo */
> +		if (params_channels(params) == 1) {
> +			value = readl(aio->cygaud->audio +
> +				aio->regs.bf_sourcech_cfg);
> +			value |= BIT(BF_SRC_CFGX_SAMPLE_CH_MODE);
> +			value &= ~BIT(BF_SRC_CFGX_BUFFER_PAIR_ENABLE);
> +			writel(value, aio->cygaud->audio +
> +				aio->regs.bf_sourcech_cfg);
> +		} else {
> +			value = readl(aio->cygaud->audio +
> +				aio->regs.bf_sourcech_cfg);
> +			value &= ~BIT(BF_SRC_CFGX_SAMPLE_CH_MODE);
> +			writel(value, aio->cygaud->audio +
> +				aio->regs.bf_sourcech_cfg);
> +		}

Either this should be a switch statement or the comment should say we
support more than stereo.  It's also not clear to me how
BUFFER_PAIR_ENABLE gets set again if we go from mono to stereo.

> +	if (!aio->is_slave) {
> +		if (aio->clk_trace.cap_clk_en)
> +			clk_prepare_enable(aio->cygaud->
> +					audio_clk[aio->pll_clk_num]);

Should check the return value of clk_prepare_enable().

> +	.playback = {
> +		.channels_min = 2,
> +		.channels_max = 2,
> +		.rates = CYGNUS_TDM_RATE | SNDRV_PCM_RATE_88200 |
> +			SNDRV_PCM_RATE_96000 | SNDRV_PCM_RATE_176400 |
> +			SNDRV_PCM_RATE_192000,
> +		.formats = SNDRV_PCM_FMTBIT_S16_LE |
> +			SNDRV_PCM_FMTBIT_S32_LE,

According to hw_params() the driver also supports S8 and S24.

> +	if (port_type == PORT_TDM) {

> +	} else { /* SPDIF case */

switch statement...

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 473 bytes --]