aac tool regressions on 7.0-RC1

[Chris Timmons]

The diff applied cleanly except for the __FBSDID().

With this new revision, I can still run batch scripts as shown below and 
get the correct output from aaccli.  Unfortunately, if you manually invoke 
aaccli and type in "open aac0", all terminal input is locked.  The only 
way to recover the process is to kill it from another window.  While this 
was going on, I did notice devd was active in top(1) output.

On the system with the hack-patch, there is no problem with interactive 
terminal I/O w/aaccli after "open aac0".  You can continue to enter 
commands interactively and see the expected output.

My 7-stable sources are from around 11am PDT today, with just the aac.c 
patched as you suggested.

-c


 	aaccli <<- EOF

 	        open aac0
 	        logfile start $LOGFILE
 	        enclosure show status
 	        logfile end
 	        close
 	        exit
 	EOF


On Wed, 2 Apr 2008, Ed Maste wrote:

> Excellent!  I've actually committed a better fix to HEAD now (aac.c
> revision 1.137).  The diffs[1] should apply cleanly to RELENG_7 and
> RELENG_6 I believe, and I plan to MFC them in a week or so.  If you have
> time to try it out on 7 before then I'd like to find out if it solves
> the problem (properly) for you.
>
> Please give aaccli a try with the driver in 7-stable and let me know if 
> you find any functionality that doesn't work.  (6-stable doesn't have 
> all of the changes yet, so I don't think it'll work there right now.)
>
> [1] http://www.freebsd.org/cgi/cvsweb.cgi/src/sys/dev/aac/aac.c.diff?r1=1.136;r2=1.137
>
> -Ed
>