[avr-libc-dev] Re: [EEPROM] patch take 3: Fix mis-feature of lacking EEPROM support for many targets / rewrite of eeprom functions

[Joerg Wunsch]

As Björn Haase wrote:

> according to your request, I have spend the time to make the
> remaining block transfers more efficient. ...

> Differing from my initial suggestion, I now agree with you that two
> dedicated write word and read word functions are useful. Take 3
> implements them.

Björn,

a big "Dankeschön" for all your work, which eventually ended up in a
true turn-key solution for the problem.  The only things I had to add
were a few bumps of the copyright dates, plus your name into the
copyright notices of the two files you've been reworking (eeprom.h,
eeprom.S).  I just committed everything to both, after verifying at
least the rudimentary EEPROM IO (on an ATmega16, and an ATmega48 which
is one of those that weren't supported before).

A similarly big "Thanks" (or rather "Spasibo") goes to everybody who
reviewed these routines, and commented on them, so Björn could
subsequently improve them.

> please have also a look at the .section directives in the eeprom.S
> file that are supposed to force the individual eeprom functions to
> be linked closely together.

I'm not much of an expert on this, but that looks OK to me.  The
linker map file shows that the .text.eeprom section is linked
contiguously.

> In take 3 I also have tried to remove whitespace changes as much as
> possible.

Great!

-- 
cheers, J"org               .-.-.   --... ...--   -.. .  DL8DTL

http://www.sax.de/~joerg/                        NIC: JW11-RIPE
Never trust an operating system you don't have sources for. ;-)