Hi David, On Sun, 6 Jan 2008 14:25:37 -0800, David Brownell wrote: > On Sunday 06 January 2008, Byron Bradley wrote: > > David, the ENOCSI comment probably applies to your at24 eeprom driver > > too. I was wondering why that error code was chosen since other than > > "No CSI structure available" I couldn't find anything about it. > > It was chosen for its uniqueness ... ideally each fault path has its > own fault code and (possibly debug-only) diagnostic. When you get > a fault code from a driver, and there's only one place that's used, > it's a lot easier to know what went wrong than if the fault code is > (over)used in many places. > > In fact it's common practice to adopt subsystem-specific conventions > about what a given errno value indicates. Otherwise, almost every > fault observed would map to a small handful ... making them useless > for fault recovery logic, and at best problematic in terms of any > diagnostic utility.
A common practice to use random error codes just to make sure they are unique? I don't think so, no. Resource not available is EBUSY, not ENOCSI, period. Please fix your driver. -- Jean Delvare _______________________________________________ i2c mailing list [email protected] http://lists.lm-sensors.org/mailman/listinfo/i2c
