On 2016-07-28 08:55, Peter Rosin wrote: > On 2016-07-26 18:51, Kachalov Anton wrote: >> I'm able to register individual slave clients on each muxed bus like the following: >> >> echo slave-24c02 0x1011 > /sys/bus/i2c/devices/i2c-9/new_device >> echo slave-24c02 0x1011 > /sys/bus/i2c/devices/i2c-10/new_device >> echo slave-24c02 0x1013 > /sys/bus/i2c/devices/i2c-11/new_device >> echo slave-24c02 0x1011 > /sys/bus/i2c/devices/i2c-12/new_device >> >> only if I do not access i2c muxed bus somewhere between adding devices, like: >> >> echo aspeed,i2c-ipmb 0x11 > /sys/bus/i2c/devices/i2c-9/new_device >> i2cdetect -y 10 >> # next command will return (-16) >> # [...] i2c i2c-10: Failed to register i2c client slave-24c02 at 0x11 (-16) >> echo slave-24c02 0x1011 > /sys/bus/i2c/devices/i2c-10/new_device >> # [...] i2c i2c-11: Failed to register i2c client slave-24c02 at 0x11 (-16) >> echo slave-24c02 0x1011 > /sys/bus/i2c/devices/i2c-11/new_device > > Is this with your patch? If so, then that makes perfect sense since > your dummy device sits on the root adapter, and you are not allowed > to create address conflicts between clients on the root adapter and > clients on downstream muxes. > > I think the problem is that when you call i2c_new_dummy, it calls > i2c_new_device without the I2C_CLIENT_SLAVE flag (which you try to > add too late). If you open code i2c_new_dummy and add this flag > before calling i2c_new_device, it might work better? No, it might not. See my reply to Wolfram... Cheers, Peter -- To unsubscribe from this list: send the line "unsubscribe linux-i2c" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
