On Tue, 17 Oct 2017 16:25:10 +0200, Takashi Sakamoto wrote:
Hi guys,
On Oct 17 2017 09:59, Kuninori Morimoto wrote:>> Some of the issues might be addressed, yes, but I'm skeptical whether
it covers all. The BE needs proper locking and refcounting that is coupled with the FE, I suppose.
So, this means, your helper patch seems OK, but DPCM side needs more hack.
In my opinion, it's better for us to take enough proofs and explanations when changing core functionalities.
Below patch for snd-dummy is to enable unbind operation to platform_device for this module. (but it's too rough to apply mainline.) This works without Iwai-san's two patches.
Let's try: $ modprobe snd-dummy $ aplay -D hw:Dummy /dev/null & $ lsmod | grep dummy $ echo -n snd_dummy.0 > /sys/bus/platform/drivers/snd_dummy/unbind $ lsmod | grep dummy $ cat /proc/asound/cards $ modprobe -r snd-dummy $ cat /proc/asound/cards
Actually, we have no addressed issues in these operations. The issue is only on ALSA SoC part.
Well, as I mentioned, a potential breakage would appear in the legacy AC97 codec binding -- if it would do any hot-unplug. But most of legacy drivers work fine without my patches just because it has only top-level unbind.
So, the top-level hot-unplug as you tested with the dummy driver works as is. The problem is only about unbinding the middle-layer component, as mentioned in the patch. It's found mostly in ASoC, but not necessarily tied only with ASoC.
But I note that even if unbind works fine to shift state of sound devices into disconnected, poll(2) call to ALSA control character devices does not return (e.g. run 'amixer events'). I don't know exactly the cause yet...
The disconnection doesn't close the device by itself (we can't), but it replaces with the dummy ops so that it never touches the driver except for closing. That is, the device is in the idle state and just accepts closing.
Takashi
======== 8< --------
From 74bb5c45f0bf36e6487538088c654b88e1efb5b5 Mon Sep 17 00:00:00 2001
From: Takashi Sakamoto o-takashi@sakamocchi.jp Date: Tue, 17 Oct 2017 17:58:47 +0900 Subject: [PATCH] ALSA: dummy: support unbind operation to shift state of sound devices to disconnected
sound/drivers/dummy.c | 39 ++++++++++++++++++++++++++++++++++++--- 1 file changed, 36 insertions(+), 3 deletions(-)
diff --git a/sound/drivers/dummy.c b/sound/drivers/dummy.c index c0939a0164a6..dcdefd2931c7 100644 --- a/sound/drivers/dummy.c +++ b/sound/drivers/dummy.c @@ -92,6 +92,7 @@ MODULE_PARM_DESC(hrtimer, "Use hrtimer as the timer source."); #endif
static struct platform_device *devices[SNDRV_CARDS]; +static struct snd_card *cards[SNDRV_CARDS];
#define MIXER_ADDR_MASTER 0 #define MIXER_ADDR_LINE 1 @@ -1134,7 +1135,24 @@ static int snd_dummy_probe(struct platform_device *devptr)
static int snd_dummy_remove(struct platform_device *devptr) {
- snd_card_free(platform_get_drvdata(devptr));
- struct snd_card *card = platform_get_drvdata(devptr);
- struct snd_device *dev;
- int i;
- for (i = 0; i < SNDRV_CARDS; ++i) {
/* Temporary for module removal. */if (devices[i] == devptr)cards[i] = card;- }
- /*
* This shifts state of an instance of sound card into disconnected, but* don't wait till all of references to instances of sound devices are* released.*/- list_for_each_entry_reverse(dev, &card->devices, list)
snd_device_disconnect(card, dev->device_data);- return 0;
}
@@ -1178,8 +1196,23 @@ static void snd_dummy_unregister_all(void) { int i;
- for (i = 0; i < ARRAY_SIZE(devices); ++i)
platform_device_unregister(devices[i]);
- for (i = 0; i < ARRAY_SIZE(devices); ++i) {
struct platform_device *devptr = devices[i];struct snd_card *card = cards[i];if (devptr == NULL)continue;if (card) {/** This can wait till the final reference to an* instance of sound card is released.*/snd_card_free(card);}platform_device_unregister(devptr);- }
- platform_driver_unregister(&snd_dummy_driver); free_fake_buffer();
}
2.11.0