On 02/05/13 14:48, Chris Clayton wrote:
[snip]
Well, after lots of hacking diagnostics into cx23885-i2c.c, I'm pretty
sure that this is a timing problem. I've eventually found that if I
insert a short delay into the top of i2c_sendbytes(), my HVR-1400
expresscard dvb-t gadget starts to work. When I run a dvb scan, it finds
all 117 services that are found using the same device on Windows 7 (and
by a nearby Samsung TV). I have no idea why the delay makes the card work.
A patch that makes this change is:
[snip]
OK, here's what I believe to be a better version of the patch:
--- linux-3.7.6/drivers/media/pci/cx23885/cx23885-i2c.c~
2013-02-01 19:46:56.000000000 +0000
+++ linux-3.7.6/drivers/media/pci/cx23885/cx23885-i2c.c 2013-02-06
11:08:31.000000000 +0000
@@ -92,6 +92,13 @@ static int i2c_sendbytes(struct i2c_adap
else
dprintk(1, "%s(msg->len=%d)\n", __func__, msg->len);
+ /* The XC3028L tuner on a WinTV-HVR-1400 fails to tune without
this */
+ if (dev->board == CX23885_BOARD_HAUPPAUGE_HVR1400) {
+ printk_once(KERN_INFO "%s - extra delay being applied for "
+ "HVR1400\n", i2c_adap->name);
+ udelay(8);
+ }
+
/* Deal with i2c probe functions with zero payload */
if (msg->len == 0) {
cx_write(bus->reg_addr, msg->addr << 25);
Signed-off-by: Chris Clayton <chris2553@xxxxxxxxxxxxxx>
It applies cleanly to 3.7.6 and to 3.80-rc6+ (pulled this morning).
Chris
--
To unsubscribe from this list: send the line "unsubscribe linux-media" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
