[i2c] [PATCH] Add a new-style driver for most I2C EEPROMs

[Wolfram Sang]

David Brownell wrote:

> That was obviously tested only at HZ=100 or with quick-ish
> EEPROMs!
Ah, okay. I was aware that msleep(1) waits not less than 10ms with 
HZ=100. I simply wondered, how to reach the timeout value for bigger 
HZ-values.

> That loop should (a) run at least a few times, maybe three;
> and (b) not exit before the timeout passes.
...or (c) return successfully as soon as possible. This is why you used 
msleep(1) and not something like msleep(AT24_EE_TIMEOUT / 2). I got it :)

 > 	timeout = ... ;
 > 	retries = 0;
 > 	while (retries++ < 3 || time_before(jiffies, timeout)) {
 > 		... try write, quit loop on success ...
 > 		msleep(1);
 > 	}
This looks good, will include it!

    Wolfram

-- 
   Dipl.-Ing. Wolfram Sang | http://www.pengutronix.de
  Pengutronix - Linux Solutions for Science and Industry