On Tue, 13 Feb 2024 01:54:11 +0100,
Wesley Cheng wrote:
>
> +static int soc_usb_get_offload_dev(struct snd_kcontrol *kcontrol,
> + struct snd_ctl_elem_value *ucontrol)
> +{
> + struct snd_soc_component *component = snd_kcontrol_chip(kcontrol);
> + struct snd_soc_usb *ctx = snd_soc_usb_find_priv_data(component->dev);
> + int ret = 0;
> +
> + mutex_lock(&ctx_mutex);
> + if (ctx && ctx->get_offload_dev)
> + ret = ctx->get_offload_dev(kcontrol, ucontrol);
> + mutex_unlock(&ctx_mutex);
It might be safer to initialize the values with -1 in case when the
callback isn't available?
> /**
> * snd_soc_usb_get_components_tag() - Retrieve SOC USB component tag
> * @playback: direction of audio stream
> @@ -157,6 +218,12 @@ EXPORT_SYMBOL_GPL(snd_soc_usb_free_port);
> */
> int snd_soc_usb_add_port(struct snd_soc_usb *usb)
> {
> + int ret;
> +
> + ret = snd_soc_usb_control_init(usb->component);
> + if (ret < 0)
> + return ret;
> +
> mutex_lock(&ctx_mutex);
> list_add_tail(&usb->list, &usb_ctx_list);
> mutex_unlock(&ctx_mutex);
We may need to remove the control element upon the driver removal,
too? In theory, you can unload the offload stuff while snd-usb-audio
is still active.
thanks,
Takashi
[Index of Archives]
[Pulseaudio]
[Linux Audio Users]
[ALSA Devel]
[Fedora Desktop]
[Fedora SELinux]
[Big List of Linux Books]
[Yosemite News]
[KDE Users]