[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20180301210428.GS12864@sirena.org.uk>
Date: Thu, 1 Mar 2018 21:04:28 +0000
From: Mark Brown <broonie@...nel.org>
To: srinivas.kandagatla@...aro.org
Cc: andy.gross@...aro.org, linux-arm-msm@...r.kernel.org,
alsa-devel@...a-project.org, david.brown@...aro.org,
robh+dt@...nel.org, mark.rutland@....com, lgirdwood@...il.com,
plai@...eaurora.org, bgoswami@...eaurora.org, perex@...ex.cz,
tiwai@...e.com, linux-soc@...r.kernel.org,
devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
linux-arm-kernel@...ts.infradead.org, rohkumar@....qualcomm.com,
spatakok@....qualcomm.com
Subject: Re: [PATCH v3 03/25] ASoC: qcom: qdsp6: Add common qdsp6 helper
functions
On Tue, Feb 13, 2018 at 04:58:15PM +0000, srinivas.kandagatla@...aro.org wrote:
> +config SND_SOC_QDSP6_COMMON
> + tristate
> + default n
> +
Ah, the other default n that had snuck in was actually in an earlier
patch in the series! I'm fairly sure this has come up with earlier
drivers you've submitted...
> +int q6dsp_map_channels(u8 ch_map[PCM_FORMAT_MAX_NUM_CHANNEL], int ch)
> +{
> + memset(ch_map, 0, PCM_FORMAT_MAX_NUM_CHANNEL);
> +
> + switch (ch) {
> + case 1:
> + ch_map[0] = PCM_CHANNEL_FC;
> + break;
> + case 2:
> + ch_map[0] = PCM_CHANNEL_FL;
> + ch_map[1] = PCM_CHANNEL_FR;
> + break;
That's some really funky indentation there...
> +static inline int q6dsp_errno(u32 error)
> +{
> + int ret = -EINVAL;
> +
> + if (error <= ARRAY_SIZE(q6dsp_err_codes))
> + ret = q6dsp_err_codes[error].lnx_err_code;
> +
> + return ret;
> +}
Is the error handling such a hot path that we need to make the lookup
functions (and the data table with the lookups) static inlines in a
header or could we just move the functions and the data into a C file?
Especially given the string lookups for the errors (which are really
nice to have) it seems wasteful.
Download attachment "signature.asc" of type "application/pgp-signature" (489 bytes)
Powered by blists - more mailing lists