Hi David, On Wed, 7 May 2008 16:25:20 -0700, David Brownell wrote: > On Wednesday 07 May 2008, Maciej W. Rozycki wrote: > > Well, I have to admit what I know in detail about I2C and how it is > > handled by Linux is what I gathered over the last weekend. The idea of > > the change, given a change had to be done, was to make the driver a bit > > more flexible. Of all the I2C RTC drivers this seems to be one of the two > > only that require *both* I2C *and* SMBus functionality in the driving > > controller at the same time. With this change in place only either of the > > two is needed. Please correct me if I am wrong. > > It checks for "I2C_FUNC_I2C | I2C_FUNC_SMBUS_BYTE_DATA" which is > admittedly odd. Because if it can do I2C, it can assuredly do > that minor subset of SMBus. (Modulo bugs in how drivers end up > reporting their capabilities.)
Not totally correct, see my reply somewhere else in this thread. > > (...) > > I think a pair of functions called i2c_smbus_read_i2c_byte_data() and > > i2c_smbus_write_i2c_byte_data() could be added to the core though. > > Too late. Those calls have been in the I2C core for a long > time now. ;) No, we don't have any function named that way. Not that I think that these names would be appropriate, though. I'd rather have i2c_smbus_read_i2c_block_emul() for example. -- Jean Delvare _______________________________________________ i2c mailing list [email protected] http://lists.lm-sensors.org/mailman/listinfo/i2c
