[PATCH] remove shm* from hdparm

[Rob Landley]

On Thursday 18 May 2006 8:40 am, Tito wrote:
> > I'll look at it in the morning, it's midnight and my laptop is STILL
> > backing up.  (The downside of having a 60 gig drive...)
>
> Hi, this patch removes shm* stuff from hdparm.c and substitutes it with ~=
>
> RESERVE_CONFIG_BUFFER(buf, TIMING_BUF_BYTES);
> mlock(buf, TIMING_BUF_BYTES);
>
> munlock(buf, TIMING_BUF_BYTES);
> RELEASE_CONFIG_BUFFER(buf);
>
> It seems to work for me.
> Please take a look at it (it's small ;-)  ) and apply if you like it.

Rummage rummage...

Removes much code, yay.  Hmmm...  But should mlock() failing be an abort, or a 
warning?  I can't do an hdparm -t /dev/hda as a normal user due to 
permissions anyway, so I guess...

As long as we've already got a quit: label, gotoing that means we only have 
one RELEASE_CONFIG_BUFFER().  (And do we _care_ if munlock() fails?  What are 
we going to do about it?  The program's probably about to exit anyway, so...)

Alright, couple small tweaks and I applied it...

Oh fun.  The dependencies for build all sources at once are screwed up.  
(build it, touch miscutils/hdparm.c, build it again.  Nothing to be done for 
_all.)

> > It compares busybox_old with busybox_unstripped.  So you need to make the
> > old version, mv busybox_unstripped busybox_old, then make the new
> > version, then make bloatcheck.
> >
> > It doesn't create busybox_old.  You have to supply that.  It'll create
> > busybox_unstripped if necessary...
>
>  Couldn't this be automated by the script
>
>  cp busybox_unstripped busybox_old

I believe I replied to that part already...

Rob
-- 
Never bet against the cheap plastic solution.