Re: hdparm script

[Simon Hürlimann <huerlisi@student.ethz.ch>]

Hi

i wrote a little script and posted a patch for the current hdparm-5.2-1 
source deb to bug #87451.

my configuration-file looks like that (had something in mind like fstab...):

# usage:
# one line per device
# <device-file> is something like '/dev/hda'
# <dma-mode> is eihter '1' or '0'. with 1=on and 0=off
# <io_32-mode> is either '1' or '0'. with 1=on and 0=off

# <device-file> <dma-mode>      <io_32-mode>

# First IDE-Disk
/dev/hda        1               1

# Second IDE-Disk
/dev/hdb        1               0

not very nice, but very easy to parse.

look at it if your interested. the script is too verbose and it supports only 
dma and io32. (multi-sector, but this is automatic get with hdparm -i.

i was able to package, and it works fine for me.
i'm ready to write the debconf scripts (config and templates...) if your 
interested. (for my script or a better one :-)

simon



Am Mittwoch, 25. September 2002 21.42 schrieb Michael Stone:
> On Wed, Sep 25, 2002 at 08:40:58PM +0100, Boris Duerner wrote:
> ># Configuration file for /etc/init.d/hdparm.
> ># Tunes hard drives on boot up.
> >#
> ># Syntax:
> >#[DRIVE]="<hdparm's parameters>"#
> >#
> ># Example:
> ># HDA="-c1 -d1 -m16"
> ># this turns 32Bit-I/O and DMA for /dev/hda on and sets
> ># multi sector count to 16.
> >#
> ># Currently only hda, hdb, hdc, hdd are supported
> ># To add support for additional devices change /etc/init.d/hdparm
> >
> >HDA=""
> >HDB=""
> >HDC=""
> >HDD=""
>
> That's pretty hideous, and also doesn't mesh well with device specific
> names (e.g., devfs). Rather than these bandaids, the proper approach
> (IMHO) is to just fix the default kernels so they turn on dma if it's
> safe.
>
> Mike Stone