On Tue, 16 Mar 2010 12:05:02 +0100, Jean Delvare wrote:
> Executive summary (as I understand it): the card that no longer works
> is a DViCO FusionHDTV7 Dual Express
> (CX23885_BOARD_DVICO_FUSIONHDTV_7_DUAL_EXP), bridge driver cx23885. It
> has 2 xc5000 chips at I2C address 0x64 (on 2 different I2C buses, of
> course), and an IR chip at 0x6b (on the first of these 2 I2C buses.)
> The latter is reported to be missing with recent dvb-v4l trees.
>
> The first thing to check is whether an ir_video I2C device is created
> or not. Look in /sys/bus/i2c/devices, list all the entries there. You
> should see two *-0064 entries for the xc5000 chips. You should also
> see, but you probably won't, one *-006b entry for the IR chip. The
> following command should let us know right away what is there:
>
> $ grep . /sys/bus/i2c/devices/*/name
>
> The ir_video device is supposed to be probed by cx23885_i2c_register().
> If it is not created, it means that the probe failed. Maybe these chips
> do not like the probe mechanism used by i2c-core (quick write) and only
> reply to reads? In that case, we'd need to use reads to detect it. The
> i2c core doesn't give us enough control to do this cleanly, but this
> could be added if the need exists. In the meantime, we can do the probe
> ourselves and instantiate the device unconditionally (by using
> i2c_new_device instead of i2c_new_probed_device).
We have been debugging over IRC with Timothy, and I have a fix which he
tested successfully. Basically, the problem is that the IR device on
his chip only replies to read commands, but when switching ir-kbd-i2c
to the standard device driver binding model in kernel 2.6.31, I changed
the probing method from quick read to quick write as a side effect.
This is why the IR device was no longer being detected. Using a quick
read again solves the issue. Here comes a fix, tested by Timothy for
the cx23885 part, untested for the cx88 part but I'd be very surprised
if cx88-based FusionHDTV did not need the exact same fix
* * * * *
IR support on FusionHDTV cards is broken since kernel 2.6.31. One side
effect of the switch to the standard binding model for IR I2C devices
was to let i2c-core do the probing instead of the ir-kbd-i2c driver.
There is a slight difference between the two probe methods: i2c-core
uses 0-byte writes, while the ir-kbd-i2c was using 0-byte reads. As
some IR I2C devices only support reads, the new probe method fails to
detect them.
For now, revert to letting the driver do the probe, using 0-byte
reads. In the future, i2c-core will be extended to let callers of
i2c_new_probed_device() provide a custom probing function.
Signed-off-by: Jean Delvare <khali@xxxxxxxxxxxx>
Tested-by: "Timothy D. Lenz" <tlenz@xxxxxxxxxx>
---
This fix applies to kernels 2.6.31 to 2.6.34. Should be sent to Linus
quickly.
drivers/media/video/cx23885/cx23885-i2c.c | 12 +++++++++++-
drivers/media/video/cx88/cx88-i2c.c | 16 +++++++++++++++-
2 files changed, 26 insertions(+), 2 deletions(-)
--- linux-2.6.34-rc1.orig/drivers/media/video/cx23885/cx23885-i2c.c 2010-02-25 09:10:33.000000000 +0100
+++ linux-2.6.34-rc1/drivers/media/video/cx23885/cx23885-i2c.c 2010-03-18 13:33:05.000000000 +0100
@@ -365,7 +365,17 @@ int cx23885_i2c_register(struct cx23885_
memset(&info, 0, sizeof(struct i2c_board_info));
strlcpy(info.type, "ir_video", I2C_NAME_SIZE);
- i2c_new_probed_device(&bus->i2c_adap, &info, addr_list);
+ /*
+ * We can't call i2c_new_probed_device() because it uses
+ * quick writes for probing and the IR receiver device only
+ * replies to reads.
+ */
+ if (i2c_smbus_xfer(&bus->i2c_adap, addr_list[0], 0,
+ I2C_SMBUS_READ, 0, I2C_SMBUS_QUICK,
+ NULL) >= 0) {
+ info.addr = addr_list[0];
+ i2c_new_device(&bus->i2c_adap, &info);
+ }
}
return bus->i2c_rc;
--- linux-2.6.34-rc1.orig/drivers/media/video/cx88/cx88-i2c.c 2010-02-25 09:08:40.000000000 +0100
+++ linux-2.6.34-rc1/drivers/media/video/cx88/cx88-i2c.c 2010-03-18 13:33:05.000000000 +0100
@@ -188,10 +188,24 @@ int cx88_i2c_init(struct cx88_core *core
0x18, 0x6b, 0x71,
I2C_CLIENT_END
};
+ const unsigned short *addrp;
memset(&info, 0, sizeof(struct i2c_board_info));
strlcpy(info.type, "ir_video", I2C_NAME_SIZE);
- i2c_new_probed_device(&core->i2c_adap, &info, addr_list);
+ /*
+ * We can't call i2c_new_probed_device() because it uses
+ * quick writes for probing and at least some R receiver
+ * devices only reply to reads.
+ */
+ for (addrp = addr_list; *addrp != I2C_CLIENT_END; addrp++) {
+ if (i2c_smbus_xfer(&core->i2c_adap, *addrp, 0,
+ I2C_SMBUS_READ, 0,
+ I2C_SMBUS_QUICK, NULL) >= 0) {
+ info.addr = *addrp;
+ i2c_new_device(&core->i2c_adap, &info);
+ break;
+ }
+ }
}
return core->i2c_rc;
}
--
Jean Delvare
--
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
