Thanks Jean. The user-space method worked great. For kernel space access, is there an easy way to get a reference to the i2c client (struct i2c_client *client) to pass to i2c_master_recv(), since I can't do open() via a kernel module? I've searched all around, but can't find such an example. I think my system would be cleaner as a kernel module (my I2C devices are ADCs, so I'd use a different analog driver over i2c for different chips in my system, and they'd all publish common analog read commands...). Thanks, Randy ----- Original Message ---- From: Jean Delvare <khali@xxxxxxxxxxxx> To: Randy <singularity_2@xxxxxxxxx> Cc: lm-sensors@xxxxxxxxxxxxxx Sent: Mon, June 6, 2011 11:48:12 PM Subject: Re: i2c device w/o registers Hi Randy, On Mon, 6 Jun 2011 17:25:10 -0700 (PDT), Randy wrote: > I have an ADC (MCP3424) that doesn't use registers/commands. > > To configure the device, you simply write the configuration to the appropriate > I2C device address. > > To read the sensor data, you simply issue a read to the device, which will > return the ADC sample and the current control register (3 or 4 bytes total, > depending on mode). > > All the commands in i2c-tools and the i2c drivers, except for the read/write > byte commands, want to take an additional register parameter. This extra param > > going over the wire messes up the chip. > > Is there any simple way to support devices that don't support registers? I > searched the archive, and found some requests for a i2c_smbus_read_word() > command, but that's about it. This question would have better been asked on the linux-i2c list, methinks. Anyway, the I2C commands you need aren't part of the SMBus command subset, so you have to use raw I2C messaging. The functions you need are i2c_master_send() and i2c_master_recv() in kernel space. In user space, you'd simply call read() and write() on the device node. I don't think there's any example of this in i2c-tools, but the i2c-dev driver supports it. -- Jean Delvare _______________________________________________ lm-sensors mailing list lm-sensors@xxxxxxxxxxxxxx http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
