Re: [x86_64] bttv: linux 2.6.16-rc1 mplayer fails to record ALSA audio and fails tune TV.

[Junichi Uekawa <dancer@netfort.gr.jp>]

Hi,

With 2.6.16-rc1, it's recording no audio, and it's recording
black-and-white noise on video. Wooo, it's getting worse.

$ uname -a
Linux dancer64 2.6.16-rc1dancer-dirty #1 Wed Jan 18 08:11:24 JST 2006 x86_64 GNU/Linux


>From the information that I've gathered so far it seems like:
1. bttv was rewritten to support v4l2
2. record via v4l2 is broken, v4l1 is not (xawtv works, mencoder broken.)
   I've heard mencoder should work with v4l1, but I have not yet confirmed.

bttv-cards.c:

        [BTTV_BOARD_GVBCTV5PCI] = {
                .name           = "IODATA GV-BCTV5/PCI",
                .video_inputs   = 3,
                .audio_inputs   = 1,
                .tuner          = 0,
                .svhs           = 2,
                .gpiomask       = 0x0f0f80,
                .muxsel         = {2, 3, 1, 0 },
                .audiomux       = {0x030000, 0x010000, 0, 0, 0x020000, 0},
                .no_msp34xx     = 1,
                .pll            = PLL_28,
                .tuner_type     = TUNER_PHILIPS_NTSC_M,
                .tuner_addr     = ADDR_UNSET,
                .radio_addr     = ADDR_UNSET,
                .audio_hook     = gvbctv5pci_audio,
                .has_radio      = 1,
        },

dmesg gives me an impression that this path is taken properly:
bttv: driver version 0.9.16 loaded
bttv: using 8 buffers with 2080k (520 pages) each for capture
bttv: Bt8xx card found (0).
bttv0: Bt878 (rev 17) at 0000:03:0b.0, irq: 22, latency: 64, mmio: 0xfdcff000
bttv0: detected: I-O Data Co. GV-BCTV5/PCI [card=81], PCI subsystem ID is 10fc:d018
bttv0: using: IODATA GV-BCTV5/PCI [card=81,autodetected]
bttv0: gpio: en=00000000, out=00000000 in=00f6f0f7 [init]
bttv0: using tuner=17
bttv0: i2c: checking for TDA9875 @ 0xb0... not found
bttv0: i2c: checking for TDA7432 @ 0x8a... not found
bttv0: i2c: checking for TDA9887 @ 0x86... not found
tuner 0-0061: chip found @ 0xc2 (bt878 #0 [sw])
tuner 0-0061: type set to 17 (Philips NTSC_M (MK2))
bttv0: registered device video0
bttv0: registered device vbi0
bttv0: registered device radio0
bttv0: PLL: 28636363 => 35468950 .. ok


v4l-info is dumped with:

        printf("inputs\n");
        for (i = 0;; i++) {
                memset(&input,0,sizeof(input));
                input.index = i;
                if (-1 == ioctl(fd,VIDIOC_ENUMINPUT,&input))
                        break;
                printf("    VIDIOC_ENUMINPUT(%d)\n",i);
                print_struct(stdout,desc_v4l2_input,&input,"",tab);
        }
        printf("\n");


which is probably handled in 
bttv-driver.c:static int bttv_common_ioctls(struct bttv *btv, unsigned int cmd, void *arg)

which unconditionally sets:
                i->audioset = 0;



> last successful one was 2.6.14-rc5.
> 
> 
> I've noticed that there is a v4l-info command, which dumps v4l2
> information, and it seems to signify that v4l2 is giving a different
> answer.
> 
> The following is a diff, which seems to signify that this difference
> exists in the two versions:
> 
> -	audioset                : 1
> +	audioset                : 0
> 


regards,
	junichi
-- 
dancer@{debian.org,netfort.gr.jp}   Debian Project