March 2017 - Alsa-devel - mailman.alsa-project.org

[alsa-devel] [PATCH] ALSA: hda/ca0132: Limit values for chip addresses to 32-bit
by Matthias Kaehlcke 01 Apr '17

01 Apr '17

With the previous unsigned long value clang generates warnings like this: sound/pci/hda/patch_ca0132.c:860:37: error: implicit conversion from 'unsigned long' to 'u32' (aka 'unsigned int') changes value from 18446744073709551615 to 4294967295 [-Werror,-Wconstant-conversion] spec->curr_chip_addx = (res < 0) ? ~0UL : chip_addx; ~ ^~~~ Signed-off-by: Matthias Kaehlcke <mka(a)chromium.org> --- sound/pci/hda/patch_ca0132.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/sound/pci/hda/patch_ca0132.c b/sound/pci/hda/patch_ca0132.c index 11b9b2f17a2e..7175e2b46fc4 100644 --- a/sound/pci/hda/patch_ca0132.c +++ b/sound/pci/hda/patch_ca0132.c @@ -857,7 +857,7 @@ static int chipio_write_address(struct hda_codec *codec, chip_addx >> 16); } - spec->curr_chip_addx = (res < 0) ? ~0UL : chip_addx; + spec->curr_chip_addx = (res < 0) ? (u32)~0U : chip_addx; return res; } @@ -882,7 +882,7 @@ static int chipio_write_data(struct hda_codec *codec, unsigned int data) /*If no error encountered, automatically increment the address as per chip behaviour*/ spec->curr_chip_addx = (res != -EIO) ? - (spec->curr_chip_addx + 4) : ~0UL; + (spec->curr_chip_addx + 4) : (u32)~0U; return res; } @@ -933,7 +933,7 @@ static int chipio_read_data(struct hda_codec *codec, unsigned int *data) /*If no error encountered, automatically increment the address as per chip behaviour*/ spec->curr_chip_addx = (res != -EIO) ? - (spec->curr_chip_addx + 4) : ~0UL; + (spec->curr_chip_addx + 4) : (u32)~0U; return res; } -- 2.12.0.367.g23dc2f6d3c-goog

2 2

Re: [alsa-devel] mono/stereo translation & number of frames per period
by Paul Stone 31 Mar '17

31 Mar '17

I set up the buffer_bytes_max to correspond with a buffer size of 4 kB. For S16_LE and stereo interleaved, that means 1024 frames in the buffer. When it goes above 512, I get a BUG warning from ALSA: [ 241.219184] ALSA sound/core/pcm_lib.c:352 BUG: pcmC3D0p:0, pos = 512, buffer size = 512, period size = 128 I guess I will have to drill down into ALSA and see where that buffer size is coming from. -Paul Stone

1 0

Re: [alsa-devel] mono/stereo translation & number of frames per period
by Paul Stone 31 Mar '17

31 Mar '17

This is becoming increasingly difficult and frustrating. I've tried to work around ALSA by setting up a different chunk size for mono vs. stereo. However, ALSA now wants the total stereo buffer size to be half as large, in accordance with the chunk size. This seems completely arbitrary to me. Why can I not specify the buffer size? I feel like I must be missing something very basic. -Paul Stone

1 0

Re: [alsa-devel] mono/stereo translation & number of frames per period
by Paul Stone 31 Mar '17

31 Mar '17

Hi, I've tried setting the runtime->hw.buffer_bytes_max in the prepare() callback, but that doesn't seem to have any effect. The copy callback continues to exceed the boundary of my monaural buffer. I've tried setting up rules based on the number of channels, but the channels interval seems to have no bearing on the actual number of channels for the substream. According to the documentation written by Takashi Iwai, it should be possible to modify the hw parameters based on the number of channels for the substream. In his example, he uses the channel minimum from the interval. I don't understand how an interval applies to a substream. I would think that an open substream would have a fixed number of channels. When I run speakertest with a single channel, I would think that is not an interval, but rather a fixed number of channels equal to '1'. So, I don't understand the purpose of accepting an interval for channels in a rules function which is intended to make changes based on the number of channels for the substream. Here is the interval structure: struct snd_interval { unsigned int min, max; unsigned int openmin:1, openmax:1, integer:1, empty:1; }; None of those fields provided to my rules function seem to contain the correct number of channels for the current substream. I am using alsa-lib-1.0.27.2-r1, which is fairly old. Best Regards, Paul Stone http://mailman.alsa-project.org/mailman/listinfo/alsa-devel

1 0

[alsa-devel] [PATCH 2/2] ALSA: timer: Info leak in snd_timer_user_tinterrupt()
by Dan Carpenter 31 Mar '17

31 Mar '17

The "r1" struct has memory holes. We clear it with memset on one path where it is used but not the other. Let's just memset it at the start of the function so it's always safe. Signed-off-by: Dan Carpenter <dan.carpenter(a)oracle.com> diff --git a/sound/core/timer.c b/sound/core/timer.c index 8b9e7943a83b..2f836ca09860 100644 --- a/sound/core/timer.c +++ b/sound/core/timer.c @@ -1277,6 +1277,7 @@ static void snd_timer_user_tinterrupt(struct snd_timer_instance *timeri, struct timespec tstamp; int prev, append = 0; + memset(&r1, 0, sizeof(r1)); memset(&tstamp, 0, sizeof(tstamp)); spin_lock(&tu->qlock); if ((tu->filter & ((1 << SNDRV_TIMER_EVENT_RESOLUTION) | @@ -1292,7 +1293,6 @@ static void snd_timer_user_tinterrupt(struct snd_timer_instance *timeri, } if ((tu->filter & (1 << SNDRV_TIMER_EVENT_RESOLUTION)) && tu->last_resolution != resolution) { - memset(&r1, 0, sizeof(r1)); r1.event = SNDRV_TIMER_EVENT_RESOLUTION; r1.tstamp = tstamp; r1.val = resolution;

2 1

[alsa-devel] [PATCH 1/2] ALSA: timer: remove some dead code
by Dan Carpenter 31 Mar '17

31 Mar '17

We just checked "id.card < 0" on the lines before so we know it's not true here. We can delete that check. Also checkpatch.pl complains about some extra curly braces so we may as well fix that while we're at it. Signed-off-by: Dan Carpenter <dan.carpenter(a)oracle.com> diff --git a/sound/core/timer.c b/sound/core/timer.c index 6d4fbc439246..8b9e7943a83b 100644 --- a/sound/core/timer.c +++ b/sound/core/timer.c @@ -1430,18 +1430,13 @@ static int snd_timer_user_next_device(struct snd_timer_id __user *_tid) if (id.card < 0) { id.card = 0; } else { - if (id.card < 0) { - id.card = 0; + if (id.device < 0) { + id.device = 0; } else { - if (id.device < 0) { - id.device = 0; - } else { - if (id.subdevice < 0) { - id.subdevice = 0; - } else { - id.subdevice++; - } - } + if (id.subdevice < 0) + id.subdevice = 0; + else + id.subdevice++; } } list_for_each(p, &snd_timer_list) {

2 1

[alsa-devel] [PATCH] ALSA: emux: stop if copy_from_user() fails
by Dan Carpenter 31 Mar '17

31 Mar '17

If we can't fill the "patch" struct because "count" is too small (it can be as low as 4 bytes) or because copy_from_user() failed, then just return instead of using unintialized data. Signed-off-by: Dan Carpenter <dan.carpenter(a)oracle.com> diff --git a/sound/synth/emux/emux_oss.c b/sound/synth/emux/emux_oss.c index ac75816ada7c..850fab4a8f3b 100644 --- a/sound/synth/emux/emux_oss.c +++ b/sound/synth/emux/emux_oss.c @@ -225,9 +225,9 @@ snd_emux_load_patch_seq_oss(struct snd_seq_oss_arg *arg, int format, else if (format == SNDRV_OSS_SOUNDFONT_PATCH) { struct soundfont_patch_info patch; if (count < (int)sizeof(patch)) - rc = -EINVAL; + return -EINVAL; if (copy_from_user(&patch, buf, sizeof(patch))) - rc = -EFAULT; + return -EFAULT; if (patch.type >= SNDRV_SFNT_LOAD_INFO && patch.type <= SNDRV_SFNT_PROBE_DATA) rc = snd_soundfont_load(emu->sflist, buf, count, SF_CLIENT_NO(p->chset.port));

2 1

[alsa-devel] [bug report] ALSA: oxfw: Add support for AV/C stream format command to get/set supported stream formation
by Dan Carpenter 31 Mar '17

31 Mar '17

Hello Takashi Sakamoto, The patch 5b59d8098d2a: "ALSA: oxfw: Add support for AV/C stream format command to get/set supported stream formation" from Dec 9, 2014, leads to the following static checker warning: sound/firewire/oxfw/oxfw-command.c:96 avc_stream_get_format() warn: check that subtract can't underflow 'err - 10' sound/firewire/oxfw/oxfw-command.c 76 /* do transaction and check buf[1-7] are the same against command */ 77 err = fcp_avc_transaction(unit, buf, 12, buf, *len, 78 BIT(1) | BIT(2) | BIT(3) | BIT(4) | BIT(5) | 79 BIT(6) | BIT(7)); 80 if ((err > 0) && (err < 10)) Assume err is either 0 or 10. 81 err = -EIO; 82 else if (buf[0] == 0x08) /* NOT IMPLEMENTED */ 83 err = -ENOSYS; 84 else if (buf[0] == 0x0a) /* REJECTED */ 85 err = -EINVAL; 86 else if (buf[0] == 0x0b) /* IN TRANSITION */ 87 err = -EAGAIN; 88 /* LIST subfunction has entry ID */ 89 else if ((subfunc == 0xc1) && (buf[10] != eid)) If "err" is 10, doesn't that mean we are reading beyond the end of the loop here? 90 err = -EIO; 91 if (err < 0) 92 goto end; 93 94 /* keep just stream format information */ 95 if (subfunc == 0xc0) { 96 memmove(buf, buf + 10, err - 10); If "err" is 0 then we are going to crash here. 97 *len = err - 10; 98 } else { 99 memmove(buf, buf + 11, err - 11); Basically if "err" is zero we're going to crash one way or another. But if err is 10 and we reach this line then we will crash as well. 100 *len = err - 11; 101 } 102 103 err = 0; 104 end: 105 return err; 106 } regards, dan carpenter

1 0

[alsa-devel] [bug report] ALSA: firewire-motu: add support for MOTU 828mk3 (FireWire/Hybrid) as a model with protocol version 3
by Dan Carpenter 31 Mar '17

31 Mar '17

Hello Takashi Sakamoto, The patch 5992e30034c4: "ALSA: firewire-motu: add support for MOTU 828mk3 (FireWire/Hybrid) as a model with protocol version 3" from Mar 22, 2017, leads to the following static checker warning: sound/firewire/motu/motu-protocol-v3.c:104 v3_get_clock_source() warn: mask and shift to zero sound/firewire/motu/motu-protocol-v3.c 90 static int v3_get_clock_source(struct snd_motu *motu, 91 enum snd_motu_clock_source *src) 92 { 93 __be32 reg; 94 u32 data; 95 unsigned int val; 96 int err; 97 98 err = snd_motu_transaction_read(motu, V3_CLOCK_STATUS_OFFSET, &reg, 99 sizeof(reg)); 100 if (err < 0) 101 return err; 102 data = be32_to_cpu(reg); 103 104 val = (data & V3_CLOCK_SOURCE_MASK) >> V3_CLOCK_SOURCE_SHIFT; 0xff >> 8 val is always zero. These are the only places we use V3_CLOCK_SOURCE_MASK and V3_CLOCK_SOURCE_SHIFT so I have no idea what it should be. 105 if (val == 0x00) { 106 *src = SND_MOTU_CLOCK_SOURCE_INTERNAL; 107 } else if (val == 0x01) { 108 *src = SND_MOTU_CLOCK_SOURCE_WORD_ON_BNC; 109 } else if (val == 0x10) { 110 *src = SND_MOTU_CLOCK_SOURCE_SPDIF_ON_COAX; 111 } else if (val == 0x18 || val == 0x19) { 112 err = snd_motu_transaction_read(motu, V3_OPT_IFACE_MODE_OFFSET, 113 &reg, sizeof(reg)); 114 if (err < 0) 115 return err; regards, dan carpenter

2 1

[alsa-devel] [PATCH] ALSA: usb-audio: Fake also USB device id when alias is given
by Takashi Iwai 31 Mar '17

31 Mar '17

Recently snd-usb-audio driver received a new option, quirk_alias, to allow user to apply the existing quirk for a different device. This works for many quirks as is, but some still need more tune-ups: namely, some quirks check the USB vendor/device IDs in various places, thus it doesn't work as long as the ID is different from the expected one. With this patch, the driver stores the aliased USB ID, so that these rest quirks per device ID are applied. The transition to use the cached USB ID was already done in the past, so what we needed now is only to overwrite chip->usb_id. Signed-off-by: Takashi Iwai <tiwai(a)suse.de> --- sound/usb/card.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/sound/usb/card.c b/sound/usb/card.c index f36cb068dad3..6640277a725b 100644 --- a/sound/usb/card.c +++ b/sound/usb/card.c @@ -332,6 +332,7 @@ static int snd_usb_audio_dev_free(struct snd_device *device) static int snd_usb_audio_create(struct usb_interface *intf, struct usb_device *dev, int idx, const struct snd_usb_audio_quirk *quirk, + unsigned int usb_id, struct snd_usb_audio **rchip) { struct snd_card *card; @@ -381,8 +382,7 @@ static int snd_usb_audio_create(struct usb_interface *intf, atomic_set(&chip->usage_count, 0); atomic_set(&chip->shutdown, 0); - chip->usb_id = USB_ID(le16_to_cpu(dev->descriptor.idVendor), - le16_to_cpu(dev->descriptor.idProduct)); + chip->usb_id = usb_id; INIT_LIST_HEAD(&chip->pcm_list); INIT_LIST_HEAD(&chip->ep_list); INIT_LIST_HEAD(&chip->midi_list); @@ -569,7 +569,7 @@ static int usb_audio_probe(struct usb_interface *intf, (vid[i] == -1 || vid[i] == USB_ID_VENDOR(id)) && (pid[i] == -1 || pid[i] == USB_ID_PRODUCT(id))) { err = snd_usb_audio_create(intf, dev, i, quirk, - &chip); + id, &chip); if (err < 0) goto __error; chip->pm_intf = intf; -- 2.11.1

2 2