From: Jean-Francois Moine <moinejf@free.fr>
To: Xiubo Li <Li.Xiubo@freescale.com>
Cc: broonie@kernel.org, lgirdwood@gmail.com, perex@perex.cz,
tiwai@suse.de, kuninori.morimoto.gx@renesas.com, jsarha@ti.com,
alsa-devel@alsa-project.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCHv2 2/2] ASoC: simple-card: Fix bug of wrong decrement DT node's refcount
Date: Mon, 1 Sep 2014 08:54:25 +0200 [thread overview]
Message-ID: <20140901085425.6553dac5@armhf> (raw)
In-Reply-To: <1409535327-11449-3-git-send-email-Li.Xiubo@freescale.com>
On Mon, 1 Sep 2014 09:35:27 +0800
Xiubo Li <Li.Xiubo@freescale.com> wrote:
> DAI links's cpu_of_node's and codec_of_node's refcounts shouldn't
> be decremented immediately at the end of the probe() fucntion.
> Because we will still use them before the audio card is removed.
Right, but your patch seems a bit complicated. See below.
>
> Signed-off-by: Xiubo Li <Li.Xiubo@freescale.com>
> ---
> sound/soc/generic/simple-card.c | 52 ++++++++++++++++++++++-------------------
> 1 file changed, 28 insertions(+), 24 deletions(-)
>
> diff --git a/sound/soc/generic/simple-card.c b/sound/soc/generic/simple-card.c
> index 9e170fe..08c5d7d 100644
> --- a/sound/soc/generic/simple-card.c
> +++ b/sound/soc/generic/simple-card.c
> @@ -307,6 +307,22 @@ dai_link_of_err:
> return ret;
> }
>
> +static inline void
> +asoc_simple_card_unref(const struct snd_soc_dai_link *dai_link,
> + int num_links)
> +{
> + struct device_node *np;
> +
> + while (num_links--) {
> + np = dai_link[num_links].cpu_of_node;
> + if (np)
> + of_node_put(np);
> + np = dai_link[num_links].codec_of_node;
> + if (np)
> + of_node_put(np);
> + }
> +}
> +
> static int asoc_simple_card_parse_of(struct device_node *node,
> struct simple_card_data *priv,
> struct device *dev,
> @@ -355,6 +371,7 @@ static int asoc_simple_card_parse_of(struct device_node *node,
> dai_props + i,
> false);
> if (ret < 0) {
> + asoc_simple_card_unref(dai_link, i + 1);
> of_node_put(np);
> return ret;
> }
> @@ -374,27 +391,6 @@ static int asoc_simple_card_parse_of(struct device_node *node,
> return 0;
> }
>
> -/* update the reference count of the devices nodes at end of probe */
> -static int asoc_simple_card_unref(struct platform_device *pdev)
> -{
> - struct snd_soc_card *card = platform_get_drvdata(pdev);
> - struct snd_soc_dai_link *dai_link;
> - struct device_node *np;
> - int num_links;
> -
> - for (num_links = 0, dai_link = card->dai_link;
> - num_links < card->num_links;
> - num_links++, dai_link++) {
> - np = (struct device_node *) dai_link->cpu_of_node;
> - if (np)
> - of_node_put(np);
> - np = (struct device_node *) dai_link->codec_of_node;
> - if (np)
> - of_node_put(np);
> - }
> - return 0;
> -}
> -
> static int asoc_simple_card_probe(struct platform_device *pdev)
> {
> struct simple_card_data *priv;
> @@ -441,7 +437,7 @@ static int asoc_simple_card_probe(struct platform_device *pdev)
> if (ret < 0) {
> if (ret != -EPROBE_DEFER)
> dev_err(dev, "parse error %d\n", ret);
> - goto err;
> + return ret;
> }
>
> } else {
> @@ -483,11 +479,18 @@ static int asoc_simple_card_probe(struct platform_device *pdev)
>
> ret = devm_snd_soc_register_card(&pdev->dev, &priv->snd_card);
>
> -err:
> - asoc_simple_card_unref(pdev);
> return ret;
> }
The main problem is there: don't unref the nodes is no error.
Why not simply:
ret = devm_snd_soc_register_card(&pdev->dev, &priv->snd_card);
+ if (ret >= 0)
+ return ret; /* success */
err:
asoc_simple_card_unref(pdev);
return ret;
}
? Then, the above stuff is not needed.
> +static int asoc_simple_card_remove(struct platform_device *pdev)
> +{
> + struct snd_soc_card *card = platform_get_drvdata(pdev);
> +
> + asoc_simple_card_unref(card->dai_link, card->num_links);
> +
> + return 0;
> +}
> +
> static const struct of_device_id asoc_simple_of_match[] = {
> { .compatible = "simple-audio-card", },
> {},
> @@ -501,6 +504,7 @@ static struct platform_driver asoc_simple_card = {
> .of_match_table = asoc_simple_of_match,
> },
> .probe = asoc_simple_card_probe,
> + .remove = asoc_simple_card_remove,
> };
>
> module_platform_driver(asoc_simple_card);
OK.
--
Ken ar c'hentañ | ** Breizh ha Linux atav! **
Jef | http://moinejf.free.fr/
next prev parent reply other threads:[~2014-09-01 6:54 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-09-01 1:35 [PATCHv2 0/2] ASoC: simple-card: Fix bug about DT node's refcount Xiubo Li
2014-09-01 1:35 ` [PATCHv2 1/2] ASoC: simple-card: Fix bug of forgetting decrement " Xiubo Li
2014-09-01 6:37 ` Jean-Francois Moine
2014-09-01 6:52 ` Li.Xiubo
2014-09-01 1:35 ` [PATCHv2 2/2] ASoC: simple-card: Fix bug of wrong " Xiubo Li
2014-09-01 6:54 ` Jean-Francois Moine [this message]
2014-09-01 7:11 ` Li.Xiubo
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20140901085425.6553dac5@armhf \
--to=moinejf@free.fr \
--cc=Li.Xiubo@freescale.com \
--cc=alsa-devel@alsa-project.org \
--cc=broonie@kernel.org \
--cc=jsarha@ti.com \
--cc=kuninori.morimoto.gx@renesas.com \
--cc=lgirdwood@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=perex@perex.cz \
--cc=tiwai@suse.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).