On Tue, 2008-06-10 at 18:36 -0500, [EMAIL PROTECTED] wrote:
> Greetings,
>
> First off I'd like to thank the developers very much for putting forth the
> effort in developing the CX18 driver for the Hauppage HVR cards. I happen
> to have a PCI interface HVR-1600, installed in an Intel i845 motherboard.
>
> I'm having trouble getting past the initial stages of getting things
> going. I've followed the steps outlined in the Cx18 wiki article
> (http://www.ivtvdriver.org/index.php/Cx18), including downloading the
> firmware and copying it to the correct location in /lib/firmware.
>
> When I load the CX18 driver, I get the following messages in dmesg:
>
> [72062.445907] cx18: Start initialization, version 1.0.0
> [72062.446390] cx18-0: Initializing card #0
> [72062.446398] cx18-0: Autodetected Hauppauge card
> [72062.446599] cx18-0: cx23418 revision 01010000 (B)
> [72062.600091] tveeprom 3-0050: Huh, no eeprom present (err=-121)?
-121 => -EREMOTEIO You have the I2C bus problem that other are
experiencing.
> [72062.600103] tveeprom 3-0050: Encountered bad packet header [00].
The all 0 byte means that one of the I2C bus slaves is stuck holding the
I2C data line low on the same bus as the EEPROM. Michael on the
ivtv-users list in the "What's next?" thread has posted positive results
by applying the attached cx18-i2c-gpio-reset3.patch (not made against
the latest - so may not apply cleanly).
But Michael had also increased the delays in the patch to 100 ms. He
also increased the mdelays() in cx18-i2c.c to 100 ms as well. 100 ms is
probably overly long, but we haven't worked down empirically what are
the smallest numbers to use.
> Corrupt or not a Hauppauge eeprom.
> [72062.600107] cx18-0: Invalid EEPROM
> [72062.600111] cx18-0: VBI is not yet supported
> [72062.954362] tuner 4-0043: chip found @ 0x86 (cx18 i2c driver #0-1)
> [72063.073754] tda9887 4-0043: creating new instance
> [72063.073764] tda9887 4-0043: tda988[5/6/7] found
> [72063.284032] tuner 4-0061: chip found @ 0xc2 (cx18 i2c driver #0-1)
> [72063.284444] cs5345 3-004c: chip found @ 0x98 (cx18 i2c driver #0-0)
> [72063.420184] cx18-0: Disabled encoder IDX device
> [72063.437246] cx18-0: Registered device video0 for encoder MPEG (2 MB)
> [72063.437723] DVB: registering new adapter (cx18)
> [72063.465382] cx18-0: Mismatch at offset 10
> [72063.465404] cx18-0: Retry loading firmware
> [72063.483676] cx18-0: Mismatch at offset 10
> [72063.483698] cx18-0: Failed to initialize on minor 0
> [72063.505208] s5h1409_readreg: readreg error (ret == -121)
> [72063.505294] cx18-0: frontend initialization failed
> [72063.519591] cx18-0: DVB failed to register
> [72063.522134] cx18-0: Registered device video32 for encoder YUV (2 MB)
> [72063.522517] cx18-0: Registered device video24 for encoder PCM audio (1 MB)
> [72063.522797] cx18-0: Registered device radio0 for encoder radio
> [72063.523245] cx18-0: Error -12 registering devices
> [72063.523873] tda9887 4-0043: destroying instance
> [72063.525311] cx18-0: Error -12 on initialization
> [72063.525541] cx18: probe of 0000:02:0a.0 failed with error -12
-12 is -ENOMEM
You're out of vmalloc space. You need a little more than 64M just to
accommodate 1 CX23418 chip since modules use vmalloc address space, and
the 64M of CX23418 register space also uses vmalloc space.
Brian Hope has reported adding a 'vmalloc=256M' option to his kernel
command line made the problem go away for his machine with 2 CX23418
chips in it.
> [72063.525786] cx18: End initialization
>
> I'm not exactly certain what that means but the crux of the problem seems
> to be that it's not identifying the eeprom correctly.
One of the I2C devices is holding the bus stuck low and hence the EEPROM
data can't be read. Somewhere down the line, that device gets unstuck
and you can start configuring the other I2C slaves like the cs5345, but
by then, it's too late for autoconfiguration. :(
Manual configuration with "modprobe cardtype=xx tuner=yy" options
should get you operational without a readable EEPROM with the right xx
and yy (I think).
Regards,
Andy
> The card is brand
> new so I don't think it's a bad eeprom, though I will admit that I have
> not installed the card in a windows PC to verify that it's working.
>
> For reference, the output of lspci -v for my HVR-1600 is:
>
> 02:0a.0 Multimedia video controller: Conexant CX23418 Single-Chip MPEG-2
> Encoder with Integrated Analog Video/Broadcast Audio Decoder
> Subsystem: Hauppauge computer works Inc. Unknown device 7444
> Flags: bus master, medium devsel, latency 64, IRQ 20
> Memory at f8000000 (32-bit, non-prefetchable) [size=64M]
> Capabilities: [44] Vital Product Data
> Capabilities: [4c] Power Management version 2
>
> The system in question is mythbuntu running with the latest versions of
> everything in their repository, and I snagged the latest version of the
> v4l-dvb drivers right before writing this email.
>
> Please let me know what I can do to assist with testing/developting the
> CX18 driver. I'm not a programmer, but I've got hardware to test with and
> am pleased to assist with development.
>
> Of course, I'd also like to watch some HDTV :) so I'd appreciate any
> advice on where to go next with getting the driver into a working state -
> as I've seen other posts from users who have gotten further than this.
>
> - Nick Bright
diff -r 1e71c0c79b3a linux/drivers/media/video/cx18/cx18-cards.c
--- a/linux/drivers/media/video/cx18/cx18-cards.c Tue Jun 03 19:12:00 2008 -0300
+++ b/linux/drivers/media/video/cx18/cx18-cards.c Thu Jun 05 06:25:33 2008 -0400
@@ -81,6 +81,10 @@ static const struct cx18_card cx18_card_
},
.gpio_init.initial_value = 0x3001,
.gpio_init.direction = 0x3001,
+ .gpio_i2c_slave_reset.active_lo_mask = 0x3001,
+ .gpio_i2c_slave_reset.active_hi_mask = 0,
+ .gpio_i2c_slave_reset.msecs_asserted = 10,
+ .gpio_i2c_slave_reset.msecs_recovery = 40,
.i2c = &cx18_i2c_std,
};
@@ -121,6 +125,10 @@ static const struct cx18_card cx18_card_
},
.gpio_init.initial_value = 0x3001,
.gpio_init.direction = 0x3001,
+ .gpio_i2c_slave_reset.active_lo_mask = 0x3001,
+ .gpio_i2c_slave_reset.active_hi_mask = 0,
+ .gpio_i2c_slave_reset.msecs_asserted = 10,
+ .gpio_i2c_slave_reset.msecs_recovery = 40,
.i2c = &cx18_i2c_std,
};
diff -r 1e71c0c79b3a linux/drivers/media/video/cx18/cx18-cards.h
--- a/linux/drivers/media/video/cx18/cx18-cards.h Tue Jun 03 19:12:00 2008 -0300
+++ b/linux/drivers/media/video/cx18/cx18-cards.h Thu Jun 05 06:25:33 2008 -0400
@@ -118,6 +118,13 @@ struct cx18_gpio_init { /* set initial G
u32 initial_value;
};
+struct cx18_gpio_i2c_slave_reset {
+ u32 active_lo_mask; /* GPIO outputs that reset i2c chips when low */
+ u32 active_hi_mask; /* GPIO outputs that reset i2c chips when high */
+ int msecs_asserted; /* time period reset must remain asserted */
+ int msecs_recovery; /* time after deassert for chips to be ready */
+};
+
struct cx18_card_tuner {
v4l2_std_id std; /* standard for which the tuner is suitable */
int tuner; /* tuner ID (from tuner.h) */
@@ -154,7 +161,8 @@ struct cx18_card {
/* GPIO card-specific settings */
u8 xceive_pin; /* XCeive tuner GPIO reset pin */
- struct cx18_gpio_init gpio_init;
+ struct cx18_gpio_init gpio_init;
+ struct cx18_gpio_i2c_slave_reset gpio_i2c_slave_reset;
struct cx18_card_tuner tuners[CX18_CARD_MAX_TUNERS];
struct cx18_card_tuner_i2c *i2c;
diff -r 1e71c0c79b3a linux/drivers/media/video/cx18/cx18-gpio.c
--- a/linux/drivers/media/video/cx18/cx18-gpio.c Tue Jun 03 19:12:00 2008 -0300
+++ b/linux/drivers/media/video/cx18/cx18-gpio.c Thu Jun 05 06:25:33 2008 -0400
@@ -57,6 +57,30 @@ static void gpio_write(struct cx18 *cx)
CX18_REG_GPIO_OUT2);
}
+static void cx18_gpio_i2c_slave_reset(struct cx18 *cx)
+{
+ const struct cx18_gpio_i2c_slave_reset *p;
+
+ p = &cx->card->gpio_i2c_slave_reset;
+
+ if ((p->active_lo_mask | p->active_hi_mask) == 0)
+ return;
+
+ /* Assuming that the masks are a subset of the bits in gpio_dir */
+
+ /* Assert */
+ cx->gpio_val =
+ (cx->gpio_val | p->active_hi_mask) & ~(p->active_lo_mask);
+ gpio_write(cx);
+ schedule_timeout_uninterruptible(msecs_to_jiffies(p->msecs_asserted));
+
+ /* Deassert */
+ cx->gpio_val =
+ (cx->gpio_val | p->active_lo_mask) & ~(p->active_hi_mask);
+ gpio_write(cx);
+ schedule_timeout_uninterruptible(msecs_to_jiffies(p->msecs_recovery));
+}
+
void cx18_gpio_init(struct cx18 *cx)
{
cx->gpio_dir = cx->card->gpio_init.direction;
@@ -75,6 +99,8 @@ void cx18_gpio_init(struct cx18 *cx)
read_reg(CX18_REG_GPIO_OUT1), read_reg(CX18_REG_GPIO_OUT2));
gpio_write(cx);
+
+ cx18_gpio_i2c_slave_reset(cx);
}
/* Xceive tuner reset function */
_______________________________________________
ivtv-devel mailing list
[email protected]
http://ivtvdriver.org/mailman/listinfo/ivtv-devel
