Re: Disk performance degradation

From: Bartlomiej Zolnierkiewicz (B.Zolnierkiewicz@elka.pw.edu.pl)
Date: Thu Jul 31 2003 - 17:17:31 EST


On Thu, 31 Jul 2003, Andrew Morton wrote:

> mru@users.sourceforge.net (Måns Rullgård) wrote:
> >
> > Mike Dresser <mdresser_l@windsormachine.com> writes:
> >
> > > Probably for reasons like that. For some reason, I can't set my ICH4
> > > based controller(ASUS P4B533) and Quantum Fireball AS40.0 to more than
> > > 255. Kernel is 2.4.21
> >
> > It appears that in 2.[56] kernels the unit for readahead is bytes,
> > rather than sectors, as used in 2.4 kernels.
>
> The ioctl which is used by
>
> blockdev --setra
>
> is still in 512-byte units.
>
> There are other backdoors such as IDE-private /proc files which can be used
> to set readahead. I'm not sure what units they use, and I don't know what
> mechanism hdparm is using to diddle readahead.

in 2.4.x:

/proc/ide/hdX/settings

breada_readahead - BLKRA{GET/SET}
file_readahead - BLKFRA{GET/SET}

both are in 1024 bytes units

in 2.6.x they are gone :-).

and hdparm is using BLKRA{GET/SET}

--
Bartlomiej

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Thu Jul 31 2003 - 22:00:52 EST