Re: Decision time.... should NATA become the default for this release?

[Sascha Wildner <saw@xxxxxxxxx>]

Matthew Dillon wrote:
>     We need to make a decision on the next release as to whether NATA
>     should become the default.  I would really like to make it the default,
>     but to do so we need some assurance that it won't break things that
>     currently work.
>
>     So I'm asking all developers running HEAD who have not already started
>     using the NATA driver to start using it instead of the ATA driver. 
>     Lets give it a good whack!
>
>     p.s. people running RELEASE should not try to use the NATA driver.

Matt,

I have a laptop at home which doesn't work with NATA, because binaries 
seemingly segfault randomly. I think several other people also have seen 
similar errors. Thomas Spanjaard suspected this might be an interrupt 
issue but so far has been unable to reproduce it because none of his 
hardware exhibits the bug (another box I have runs fine with NATA).

It sounds similar to what was reported here:

http://leaf.dragonflybsd.org/mailarchive/kernel/2007-02/msg00122.html

. ..except that it even gets to the login: prompt sometimes (with lots of 
stuff crashing on the way). LiveCDs with NATA support compiled in work 
fine, btw.

I'll upgrade the laptop later tonight to see if it still has problems. 
If it does, I'll be happy to assist in bugfixing if you have any idea 
how we could find out what's wrong.

I'll report how latest HEAD works.

Sascha

--
http://yoyodyne.ath.cx