When ASoC driver is unbound dynamically during its operation (i.e. a
kind of hot-unplug), we may hit Oops due to the resource access after
the release by a delayed work, something like:
Unable to handle kernel paging request at virtual address
dead000000000220
....
PC is at soc_dapm_dai_stream_event.isra.14+0x20/0xd0
LR is at snd_soc_dapm_stream_event+0x74/0xa8
....
[<
ffff000008715610>] soc_dapm_dai_stream_event.isra.14+0x20/0xd0
[<
ffff00000871989c>] snd_soc_dapm_stream_event+0x74/0xa8
[<
ffff00000871b23c>] close_delayed_work+0x3c/0x50
[<
ffff0000080bbd6c>] process_one_work+0x1ac/0x318
[<
ffff0000080bbf20>] worker_thread+0x48/0x420
[<
ffff0000080c201c>] kthread+0xfc/0x128
[<
ffff0000080842f0>] ret_from_fork+0x10/0x18
For fixing the race, this patch adds a sync-point in pcm private_free
callback to finish the delayed work before actually releasing the
resources.
Reported-by: Hiep Cao Minh <cm-hiep@jinso.co.jp>
Reported-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Tested-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Mark Brown <broonie@kernel.org>
return ret;
}
+static void soc_pcm_private_free(struct snd_pcm *pcm)
+{
+ struct snd_soc_pcm_runtime *rtd = pcm->private_data;
+ struct snd_soc_platform *platform = rtd->platform;
+
+ /* need to sync the delayed work before releasing resources */
+ flush_delayed_work(&rtd->delayed_work);
+ if (platform->driver->pcm_free)
+ platform->driver->pcm_free(pcm);
+}
+
/* create a new pcm */
int soc_new_pcm(struct snd_soc_pcm_runtime *rtd, int num)
{
}
}
- pcm->private_free = platform->driver->pcm_free;
+ pcm->private_free = soc_pcm_private_free;
out:
dev_info(rtd->card->dev, "%s <-> %s mapping ok\n",
(rtd->num_codecs > 1) ? "multicodec" : rtd->codec_dai->name,