[alsa-devel] [PATCH v2 0/2] ALSA: usb-audio: fix PCM device order
Hi,
this is a rewrite of the old USB-audio fix patch I obviously overlooked to adapt to 5.3 kernel with a preliminary cleanup patch.
Totally untested, so any test / feedback would be appreciated.
thanks,
Takashi
===
Alexander Tsoy (1): ALSA: usb-audio: fix PCM device order
Takashi Iwai (1): ALSA: usb-audio: Unify audioformat release code
sound/usb/stream.c | 64 +++++++++++++++++++++++++++++++++++++++--------------- 1 file changed, 47 insertions(+), 17 deletions(-)
There are many open code for releasing audioformat object. Provide a unified helper and call it from the all places.
Only a cleanup, no functional changes.
Signed-off-by: Takashi Iwai tiwai@suse.de --- sound/usb/stream.c | 30 ++++++++++++++---------------- 1 file changed, 14 insertions(+), 16 deletions(-)
diff --git a/sound/usb/stream.c b/sound/usb/stream.c index 7ee9d17d0143..963d425004f8 100644 --- a/sound/usb/stream.c +++ b/sound/usb/stream.c @@ -28,6 +28,14 @@ #include "power.h" #include "media.h"
+static void audioformat_free(struct audioformat *fp) +{ + list_del(&fp->list); /* unlink for avoiding double-free */ + kfree(fp->rate_table); + kfree(fp->chmap); + kfree(fp); +} + /* * free a substream */ @@ -37,11 +45,8 @@ static void free_substream(struct snd_usb_substream *subs)
if (!subs->num_formats) return; /* not initialized */ - list_for_each_entry_safe(fp, n, &subs->fmt_list, list) { - kfree(fp->rate_table); - kfree(fp->chmap); - kfree(fp); - } + list_for_each_entry_safe(fp, n, &subs->fmt_list, list) + audioformat_free(fp); kfree(subs->rate_list.list); kfree(subs->str_pd); snd_media_stream_delete(subs); @@ -832,8 +837,7 @@ snd_usb_get_audioformat_uac12(struct snd_usb_audio *chip, /* ok, let's parse further... */ if (snd_usb_parse_audio_format(chip, fp, format, fmt, stream) < 0) { - kfree(fp->rate_table); - kfree(fp); + audioformat_free(fp); return NULL; }
@@ -1043,8 +1047,7 @@ snd_usb_get_audioformat_uac3(struct snd_usb_audio *chip,
pd = kzalloc(sizeof(*pd), GFP_KERNEL); if (!pd) { - kfree(fp->rate_table); - kfree(fp); + audioformat_free(fp); return NULL; } pd->pd_id = (stream == SNDRV_PCM_STREAM_PLAYBACK) ? @@ -1063,9 +1066,7 @@ snd_usb_get_audioformat_uac3(struct snd_usb_audio *chip, /* ok, let's parse further... */ if (snd_usb_parse_audio_format_v3(chip, fp, as, stream) < 0) { kfree(pd); - kfree(fp->chmap); - kfree(fp->rate_table); - kfree(fp); + audioformat_free(fp); return NULL; } } @@ -1184,11 +1185,8 @@ int snd_usb_parse_audio_interface(struct snd_usb_audio *chip, int iface_no) err = snd_usb_add_audio_stream(chip, stream, fp);
if (err < 0) { - list_del(&fp->list); /* unlink for avoiding double-free */ + audioformat_free(fp); kfree(pd); - kfree(fp->rate_table); - kfree(fp->chmap); - kfree(fp); return err; } /* try to set the interface... */
From: Alexander Tsoy alexander@tsoy.me
Some cards have alternate setting with non-PCM format as the first altsetting in the interface descriptors. This confuses userspace, since alsa-lib uses device 0 by default. So lets parse interfaces in two steps: 1. Parse altsettings with PCM formats. 2. Parse altsettings with non-PCM formats.
This fixes at least following cards: - Audinst HUD-mx2 - Audinst HUD-mini
[ Adapted 5.3 kernel by tiwai ]
Signed-off-by: Alexander Tsoy alexander@tsoy.me Signed-off-by: Takashi Iwai tiwai@suse.de --- sound/usb/stream.c | 34 +++++++++++++++++++++++++++++++++- 1 file changed, 33 insertions(+), 1 deletion(-)
diff --git a/sound/usb/stream.c b/sound/usb/stream.c index 963d425004f8..fc3e9fcfbc38 100644 --- a/sound/usb/stream.c +++ b/sound/usb/stream.c @@ -1077,7 +1077,9 @@ snd_usb_get_audioformat_uac3(struct snd_usb_audio *chip, return fp; }
-int snd_usb_parse_audio_interface(struct snd_usb_audio *chip, int iface_no) +static int __snd_usb_parse_audio_interface(struct snd_usb_audio *chip, + int iface_no, + bool *has_non_pcm, bool non_pcm) { struct usb_device *dev; struct usb_interface *iface; @@ -1178,6 +1180,16 @@ int snd_usb_parse_audio_interface(struct snd_usb_audio *chip, int iface_no) else if (IS_ERR(fp)) return PTR_ERR(fp);
+ if (fp->fmt_type != UAC_FORMAT_TYPE_I) + *has_non_pcm = true; + if ((fp->fmt_type == UAC_FORMAT_TYPE_I) == non_pcm) { + audioformat_free(fp); + kfree(pd); + fp = NULL; + pd = NULL; + continue; + } + dev_dbg(&dev->dev, "%u:%d: add audio endpoint %#x\n", iface_no, altno, fp->endpoint); if (protocol == UAC_VERSION_3) err = snd_usb_add_audio_stream_v3(chip, stream, fp, pd); @@ -1197,3 +1209,23 @@ int snd_usb_parse_audio_interface(struct snd_usb_audio *chip, int iface_no) return 0; }
+int snd_usb_parse_audio_interface(struct snd_usb_audio *chip, int iface_no) +{ + int err; + bool has_non_pcm = false; + + /* parse PCM formats */ + err = __snd_usb_parse_audio_interface(chip, iface_no, &has_non_pcm, false); + if (err < 0) + return err; + + if (has_non_pcm) { + /* parse non-PCM formats */ + err = __snd_usb_parse_audio_interface(chip, iface_no, &has_non_pcm, true); + if (err < 0) + return err; + } + + return 0; +} +
On Sat, 2019-07-27 at 11:30 +0200, Takashi Iwai wrote:
Hi,
this is a rewrite of the old USB-audio fix patch I obviously overlooked to adapt to 5.3 kernel with a preliminary cleanup patch.
Totally untested, so any test / feedback would be appreciated.
Thanks for the patches! Unfortunately I can't test them, and the reporter who prompted me to ping apparently isn't going to test them either. Hopefully the patches are still relevant to Alexander so maybe he could test them.
В Вс, 28/07/2019 в 12:56 +0300, Tanu Kaskinen пишет:
On Sat, 2019-07-27 at 11:30 +0200, Takashi Iwai wrote:
Hi,
this is a rewrite of the old USB-audio fix patch I obviously overlooked to adapt to 5.3 kernel with a preliminary cleanup patch.
Totally untested, so any test / feedback would be appreciated.
Thanks for the patches! Unfortunately I can't test them, and the reporter who prompted me to ping apparently isn't going to test them either. Hopefully the patches are still relevant to Alexander so maybe he could test them.
Sorry for the late reply, I messed up my mail filters. HUD-mx2 now works as expected without workarounds in alsa or pulse. Tested with 5.4.x. Thanks!
On Mon, 30 Dec 2019 19:01:43 +0100, Alexander Tsoy wrote:
В Вс, 28/07/2019 в 12:56 +0300, Tanu Kaskinen пишет:
On Sat, 2019-07-27 at 11:30 +0200, Takashi Iwai wrote:
Hi,
this is a rewrite of the old USB-audio fix patch I obviously overlooked to adapt to 5.3 kernel with a preliminary cleanup patch.
Totally untested, so any test / feedback would be appreciated.
Thanks for the patches! Unfortunately I can't test them, and the reporter who prompted me to ping apparently isn't going to test them either. Hopefully the patches are still relevant to Alexander so maybe he could test them.
Sorry for the late reply, I messed up my mail filters. HUD-mx2 now works as expected without workarounds in alsa or pulse. Tested with 5.4.x. Thanks!
Good to hear.
So far there is no real regression report, so it seems working for other devices, too.
thanks,
Takashi
participants (3)
-
Alexander Tsoy -
Takashi Iwai -
Tanu Kaskinen