[alsa-devel] [RFC 1/4] ASoc: Use ref_count for soc DAI & component
Mark Brown
broonie at kernel.org
Mon Feb 15 14:59:41 CET 2016
On Mon, Feb 15, 2016 at 05:49:29PM +0530, Vaibhav Agarwal wrote:
> This is preperation to allow dynamic DAI link insertion &
> removal.
Which there is no proposed user of here...
> Currently, DAI links are added/removed once during soc-card
> instatiate & removal. Thus, irrespective of usage count by multiple
> DAI links, DAIs and components are probed or removed only once
> while maintaining 'probed' flag.
> However, in case of dynamic DAI link insertion/removal we need to
> ensure DAI/components are not unnecessarily probed multiple & not
> removed mistakenly while in use by any other existing DAI link.
> Thus, ref_count is used to maintain their usage count.
Please use normal changelog formatting - either consistent line lengths
or blank lines between paragraphs depending on what you're trying to
accomplish here.
> - if (!dai->probed && dai->driver->probe_order == order) {
> - if (dai->driver->probe) {
> - ret = dai->driver->probe(dai);
> - if (ret < 0) {
> - dev_err(dai->dev,
> - "ASoC: failed to probe DAI %s: %d\n",
> - dai->name, ret);
> - return ret;
> + if (dai->driver->probe_order == order) {
> + if (!dai->probed) {
> + if (dai->driver->probe) {
> + ret = dai->driver->probe(dai);
> + if (ret < 0) {
> + dev_err(dai->dev,
> + "ASoC: failed to probe DAI %s: %d\n",
> + dai->name, ret);
> + return ret;
> + }
> }
> - }
>
> - dai->probed = 1;
> + dai->probed = 1;
> + }
> + dai->ref_count++;
> }
I don't understand this at all - why the complete reindent and change of
condition, don't we just increase the refcount when we flag as probed
(and why do we still have the probed flag after this)?
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: not available
URL: <http://mailman.alsa-project.org/pipermail/alsa-devel/attachments/20160215/4fc568a3/attachment.sig>
More information about the Alsa-devel
mailing list